Investigation of Coiled-Coil Interactions Between Proteins of the Spindle Pole Body

Investigation of Coiled-Coil Interactions between Proteins of the Spindle Pole Body

Nora Zizlsperger

B.S. Biochemistry Simmons College, 2000

SUBMITTED TO THE DEPARTMENT OF BIOLOGY IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY AT THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY

FEBRUARY 2010

Signature of Author………………………………………………………………………………… Nora Zizlsperger Department of Biology November 19, 2009

Certified by………………………………………………………………………………………… Amy E. Keating Associate Professor of Biology Thesis Supervisor

Accepted by………………………………………………………………………………………... Steven P. Bell Chairman, Department Committee on Graduate Students

Investigation of Coiled-Coil Interactions between Proteins of the Spindle Pole Body

by Nora Zizlsperger

Submitted to the Department of Biology on November 19, 2009 in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Biology

ABSTRACT

The spindle pole body (SPB) is a large multi-protein complex that organizes microtubules in yeast. Through its function of nucleating and anchoring microtubules, the SPB is essential for cell viability. High-resolution structures of the SPB have not been achieved using x-ray crystallography, due to its low copy number, large size, heterogeneous composition, and association with the nuclear membrane. However, structural information may be deciphered through a variety of other techniques. Cryo-electron microscopy images have provided a low- resolution model of the SPB. Experiments testing which proteins interact provide additional structural data, although in most cases, it is not known precisely how these interactions occur. Interestingly, a large proportion of SPB proteins are predicted to contain one or more coiled coils. The coiled coil is a common protein-protein interaction domain, consisting of two or more supercoiled α-helices. The high frequency of coiled coils predicted in SPB proteins suggests that this structure may be important for establishing the overall architecture of the complex. This thesis describes work towards determining whether coiled coils form interactions within or between SPB proteins. All coiled-coil regions predicted in SPB proteins were produced and tested for interactions as individual peptides, taking advantage of the often-observed ability of coiled coils to fold and interact cooperatively, isolated from the rest of the protein. Many self- associating coiled coils and several hetero-associating coiled coils were identified, and structural features of several complexes were determined. The results suggest that several SPB coiled coils form supports between layers of the SPB and other coiled coils mediate homo- and hetero- associations within layers or sub-complexes. The coiled-coil interaction data were incorporated with other structural information described in the literature to generate a model of the spindle pole body structure.

Thesis Supervisor: Amy E. Keating Title: Associate Professor of Biology

Table of Contents

List of Figures ...... 7 List of Tables ...... 8

CHAPTER 1 Introduction ...... 9 Structure determination of large protein complexes through integration of diverse data ...... 9 Methods for obtaining structural information about complexes ...... 10 High-resolution structural studies ...... 10 Computational structure prediction and annotation ...... 10 Electron microscopy ...... 11 Biochemical identification of components and interactions ...... 11 Models of complex structures ...... 12 Progress towards structural models of particular large protein complexes ...... 13 The proteasome ...... 14 The kinetochore ...... 14 The nuclear pore complex ...... 15 The spindle pole body ...... 16 Coiled coils – predicted structural features in the SPB ...... 18 Summary of thesis ...... 21

CHAPTER 2 Analysis of coiled-coil interactions between core proteins of the spindle pole body ...... 23 ABSTRACT ...... 23 INTRODUCTION ...... 24 MATERIAL AND METHODS ...... 25 Coiled-coil prediction ...... 25 Domain annotation ...... 29 Protein production ...... 29 Circular dichroism spectroscopy ...... 30 Fluorescence resonance energy transfer (FRET) assay ...... 30 Analytical ultracentrifugation ...... 31 Crosslinking ...... 31 Crystallization ...... 31 Structure determination ...... 31 RESUTS ...... 32 Many coiled coils, but few other domains, are predicted in core SPB protein sequences. ... 32 Several core SPB coiled coils self-associate ...... 33 Several core SPB coiled coils hetero-associate...... 42 Characterization of core SPB coiled-coil interactions ...... 44 Crystal structure of Spc42_1 ...... 47 DISCUSSION ...... 50

CHAPTER 3 Specific coiled-coil interactions contribute to a global model of the structure of the spindle pole body ...... 53 ABSTRACT ...... 53 INTRODUCTION ...... 53 MATERIALS AND METHODS ...... 55 Coiled-coil prediction and production...... 55 Circular dichroism spectroscopy, analytical ultracentrifugation, and crosslinking experiments ...... 56 Fluorescence resonance energy transfer (FRET) assay ...... 56 Fluorescence anisotropy (FA) assay ...... 56 Domain annotation...... 57 RESULTS ...... 57 Many coiled coils are predicted in membrane-associated and half-bridge SPB proteins. .... 57 Many coiled coils from SPB membrane-associated proteins self-associate...... 57 Characterization of homo-oligomerizing coiled coils in membrane-associated proteins ..... 62 Coiled coils from Bbp1 and Mps2 hetero-associate...... 64 Putative coiled coils were not observed to participate in multi-component interactions...... 67 Computational annotation predicts few structural features in SPB proteins besides coiled coils...... 70 DISCUSSION ...... 71 Overall structure ...... 74 γ-tubulin complex ...... 75 Core proteins ...... 76 Membrane-associated proteins ...... 80 Half-bridge proteins ...... 81 Summary and future directions ...... 82

CHAPTER 4 Conclusions and Future Directions ...... 85 Conclusions ...... 85 Future Directions ...... 87 Additional characterization of SPB coiled coils ...... 87 Coiled coils in other complexes ...... 90

REFERNCES ...... 92

APPENDIX ...... 103

List of Figures

Figure 1.1 Cryo-EM image of SPB ...... 16

Figure 2.1 A schematic of the SPB core complex ...... 28 Figure 2.S1 Paircoil2 score for Cnm67...... 28 Figure 2.2 CD spectra for core SPB coiled coils ...... 35 Figure 2.3 FRET assay of pair-wise interactions ...... 36 Figure 2.S2 FRET assay for mixtures of control coiled coils ...... 37 Figure 2.S3 FRET assay for mixtures of SPB coiled coils under different conditions ...... 38 Figure 2.S4 Orientation test for SPB coiled coils under different conditions ...... 40 Figure 2.S5 Hetero-associations evaluated by crosslinking ...... 43 Figure 2.4 Characterization by analytical ultracentrifugation ...... 45 Figure 2.S6 Concentration dependence of Spc72_1 ...... 46 Figure 2.5 Crystal structure of Spc42_1 ...... 49

Figure 3.1 CD spectra for coiled coils from SPB proteins ...... 58 Figure 3.2 FRET assay of pairwise interactions ...... 60 Figure 3.3 Oligomerization assessed by analytical ultracentrifugation ...... 62 Figure 3.4 Confirmation of hetero-association between Bbp1_2 and Mps2_2 ...... 64 Figure 3.5 FRET assay for multi-component interactions ...... 67 Figure 3.6 Fluorescence anisotropy (FA) assay for multi-component interactions ...... 68 Figure 3.7 Cross-section of the SPB ...... 70 Figure 3.8 Structures, models, and interactions of SPB proteins ...... 71

List of Tables

Table 2.1 Predicted coiled-coil regions of core SPB protein sequences ...... 26 Table 2.S1 Predicted coiled-coil regions of core SPB protein sequences ...... 27 Table 2.2 Summary of biophysical analyses of coiled coils of the SPB core proteins ...... 34 Table 2.3 Data processing and refinement statistics for Spc42_1 ...... 48

Table 3.1 Predicted coiled-coil regions of membrane-associated and half-bridge SPB protein sequences and summary of CD experiments ...... 58 Table 3.2 Oligomerization of coiled coils from SPB membrane-associated proteins ...... 61

Appendix Predicted coiled-coil regions of SPB proteins – DNA and protein sequences…...... 92

CHAPTER 1

Introduction

Structure determination of large protein complexes through integration of diverse data Many important cellular functions are performed by proteins, which often work together in complexes. The complexes vary in size, morphology, and heterogeneity, which affect the ease of studying them. They can be as simple as 2-10 copies of a polypeptide, arranged to influence one another, for example hemoglobin. Or they can be as intricate as hundreds to thousands of different proteins, present in varied proportions, that interact dynamically for a particular purpose. Examples of these more elaborate complexes include the transcriptional machinery, the nuclear pore complex, and microtubule organizing centers (MTOC). Three-dimensional (3D) structural information is crucial to understanding the molecular details of complex assembly and function. Ideally, a high-resolution x-ray crystal structure of an entire complex is examined for insight, but structural data may be difficult or impossible to obtain with current technology if the complex is too large, too hard to purify to homogeneity, and/or too low in symmetry. Alternatively, an intractable complex can be investigated with a variety of complementary methods, such as those described below, which provide indirect or low-resolution information about the structure. The data from these methods can be combined to generate a structural model of the complex (1). This thesis describes my efforts to understand the structure of a particular large complex, the spindle pole body (SPB). I contribute indirect structural data through the investigation of coiled-coil interactions between SPB proteins, and I integrate these data with other structural information to build a model of the SPB structure. In this Introduction, I review various methods that can impart structural details for large complexes. To illustrate how these methods can be used to construct useful structural models of large complexes, I provide three examples, in varying stages of progress. I conclude with an overture of the SPB and the numerous coiled coils in its proteins.

Methods for obtaining structural information about complexes A range of methods are available to produce different types of structural information. Here I describe each method’s spatial features, resolution, strengths, and limitations. While many of these methods can be used to address a variety of biological questions, I focus on their utility for understanding the structure of large complexes.

High-resolution structural studies X-ray crystallography is the most popular technique for obtaining high-resolution structures of proteins and complexes. X-ray diffraction data from crystallized protein complexes are used to optimize atomic-level models that fit the data. Many labs and consortia have solved structures of proteins or complexes; over 60,000 structures have been deposited (9/09) in the Protein Data Bank (PDB) from > 22,000 proteins (> 90% similar) (2). Despite the large number of successfully determined structures, many complexes have eluded structure determination. Large quantities of complexes must be purified, and the samples must be able to be concentrated and crystallized. Strategies for obtaining crystals include over-expressing the complex recombinantly, using orthologous complexes from other organisms, and extensively searching for solution conditions that allow crystallization. The crystals must then diffract X-rays to produce a sufficient number of well-spaced reflections that can be converted to a set of structure factor amplitudes. Phase information can be experimentally determined from heavy-metal derivatives or acquired from prior, related structures. Computer programs are used to convert this information to an image of electron density and to build a molecular model that fits the image. In cases where an entire complex is intractable, this technique can still be effective in structure determination of individual proteins from the complex. Individual proteins, or small sets of interacting proteins, may be easier and more competent to purify and form diffractable crystals than the entire complex. If a protein is smaller than 100 kDa, its structure may be determined by NMR spectroscopy, avoiding the need to crystallize the protein. These structures can give atomic-level details for the individual proteins, and overall shapes that can be fit into the complex.

Computational structure prediction and annotation In the absence of an experimentally-derived structure, a protein’s structural features may be predicted using various sequence-based methods (3-7). The choice and success of the method 10

depend on how similar the sequence of the target protein is to protein sequences with known structures. Strong similarity often correlates to more reliable and accurate models. If a protein sequence is > 30% identical to a protein sequence with known structure, comparative/homology modeling may be used to make models with up to 1-2 Å deviations. For sequences with no close homologs, general fold assignments may be achieved by de novo prediction or threading methods using libraries of known folds or structural fragments. Resulting models vary in accuracy from structures with errors just in loop regions to entirely incorrect fold designations. Structural features, including secondary structure, transmembrane segments, disordered regions, coiled coils and other well-characterized motifs, can often be detected by sequence with reliable accuracy. Combining multiple methods can increase confidence in prediction of protein structure.

Electron microscopy Low or moderate-resolution structures of complexes may be obtained through different types of cryo-electron microscopy (cryo-EM) (8). Electron crystallography can produce structures with resolutions as high as 2 Å and involves electron diffraction and imaging of 2D crystals of complexes to obtain models. If a complex does not crystallize, single particle cryo- EM can be used to image large numbers of purified complexes in different orientations. In this method, the 2D images of individual complexes are aligned and combined computationally to construct a 3D model of the complex. The resolution depends on the quality and quantity of complexes available (more is better), the quantity of images averaged (more is better), and the accuracy of aligning the images. If a complex is too heterogeneous or hard to purify, electron tomography can be used to image a single complex tilted in multiple orientations to provide 3D data. However, inferior images are obtained because the electron dose is very low to prevent destruction of the samples through multiple images. In addition, images are missing at some orientations due to the physical constraints of the tilt angle. These problems can be overcome with symmetry in the complex.

Biochemical identification of components and interactions Early efforts to characterize a particular complex involved identification of proteins through biochemical purifications of the complex or genes through genetic screens that would perturb the complex. This “parts” list was often confirmed by microscopy studies co-localizing

tagged versions of each protein in vivo. If the complex is large enough to be observed by conventional EM, proteins can be localized to particular regions of the complex. These proteins may be recombinantly expressed individually and studied in vitro. Knowing the identity and even the structure of the proteins within a complex is not sufficient to build a model of the complex. Investigation of which components interact facilitates placement of the proteins within a complex. Interaction assays are also often used to identify members of complexes. The most common assays for detecting protein-protein interactions are yeast two-hybrid and affinity purification/co-immunoprecipitation, both of which have been adapted for high-throughput screens (9). Particular interactions may not be detected by both assays, likely because each assay is biased towards detecting certain types of interactions. In general, yeast two-hybrid assays find transient, binary, direct interactions whereas affinity purification finds stable, multi-component, indirect interactions. Complementary assays, such as FRET, protein complementation assays, or phage/yeast display, can identify or confirm interactions and help characterize the interactions further (10). Characterization of interactions can improve resolution about how proteins fit within a complex. In particular, localizing the binding interface between the interacting proteins can help orient proteins relative to each other and may indicate their structure. This can be done by testing fragments of the proteins to narrow down the region required for interaction. Fragments can be defined using domain boundaries or more ad hoc constructs. Interactions between proteins are often mediated by conserved domains. Many recurring interaction domains have a shared structure and sequence, but also have elements of specificity to ensure that the correct proteins come together. Some interactions consist of a globular domain from one protein bound to a linear peptide motif from another protein. Other interactions feature much larger, more hydrophobic interfaces between two domains. Specific residues within the binding interface can be probed with site-directed mutagenesis and crosslinking/mass-spectrometry experiments.

Models of complex structures Structural models are often proposed to encompass and summarize available data. They are useful in understanding what the data represent and determining what new experiments are needed to further investigate a complex. Models come in many forms with varying levels of resolution and completeness. Low-resolution models based on EM images may just describe the

overall shape and differential density of a complex. Ball-and-stick models may represent constituent proteins as balls and the interactions between them as sticks. Cartoon models may portray the complex and proteins as shapes with relative size constraints. The proteins may be localized within the complex based on microscopy or interaction studies. At higher resolution, crystal structures of complex proteins may be fit into cryo-EM images with rigid-body-docking algorithms. Higher accuracy may be achieved by introducing flexibility through dividing large proteins into its domains or using molecular dynamics (11, 12). Incorporation of different types of data into a model often requires the insight and conception of investigators. This may be very difficult when the data are in different forms, are at different levels of accuracy and resolution, and/or only describe part of the complex. Recently, rather than doing this incorporation by hand, the Sali and Rout labs generated a model of the nuclear pore complex computationally (13, 14). Experimental data were converted into spatial restraints with ranges for resolution and accuracy. A scoring function was established to maximize the number of satisfied restraints and to minimize spatial overlap. The model was optimized in an iterative process of moving 3D representations of proteins randomly in space and scoring the configuration. A single model of the nuclear pore complex was found to satisfy the restraints. This computational approach is a powerful framework to build structural models of complexes. It may not be appropriate for all complexes, because a certain amount of data is needed to begin the process. Surveying the literature for structural information about a complex is useful in determining what other data are required to build a comprehensive model. Even with enough data to start building a model, the algorithm may find multiple models consistent with the data available, but may suggest additional experiments to rule out one or more of them. Even if a single model accommodates all of the data, more structural information can be incorporated to yield a higher-resolution structure.

Progress towards structural models of particular large protein complexes The focus of my thesis is on a particular protein complex, the spindle pole body. To build a model of its structure, this complex requires integration of data from experiments such as those described above. Exploration of what types of data to use, how to integrate the data, and how to build a model can be accomplished by examining efforts to understand the structure of other

large protein complexes. Here I present the current structural models for three complexes, to provide ideas for how to build a structural model of the spindle pole body.

The proteasome The eukaryotic 26S proteasome is a 2.5 MDa barrel-shaped complex and functions to degrade ubiquitin-tagged proteins (15, 16). It can be divided into two major parts: the 19S regulatory particle (RP) that recognizes and unfolds ubiquitin-tagged proteins, and the 20S catalytic core particle (CP) that degrades the proteins. The 20S CP was tractable for structure determination by x-ray crystallography (17, 18). The structure shows four heptameric rings – two outer rings of alpha1-7 subunits and two inner rings of beta1-7 subunits. The entire 26S complex and the 19S RP have resisted structural determination, due to the instability of the complex during purification, and the dynamic, heterogeneous nature of the 19S RP. Recent single particle cryo-EM experiments of the entire 26S proteasome from Drosophila and humans provide a medium-resolution (~20 Å) model, into which the 20S CP crystal structure fits well (19-21). The 19S RP connects to either end of the 20S CP and can be divided into two major parts: the base and the lid. A model of the base was built computationally by fitting homology models for each of its nine subunits into the cryo-EM map and using interaction constraints from biochemical experiments (20). A model of the lid has only been proposed in cartoon shapes of its nine subunits, but the arrangement of the subunits has been mapped based on interaction experiments. These included yeast two-hybrid analysis, partial complex assembly in lid protein mutants, and mass spectrometry of the intact lid complex (16). Additional structural characterization of the lid part of the 19S will help complete a structural model of the 26S proteasome.

The kinetochore Kinetochores connect chromosomes at centromeres to the positive end of microtubules (22, 23). Low-resolution cryo-EM images have revealed the structure of the kinetochore as a thick short cylinder of two electron-dense layers, named the inner and outer kinetochores (24). A higher-resolution structure of the kinetochore by x-ray crystallography or even single-particle cryo-EM is technically difficult to obtain, since the complex is highly dynamic and challenging to purify. However, over 80 different proteins have been identified as kinetochore components or regulators and many of these proteins have been shown to co-purify in smaller stable sub- complexes (22-27). The sub-complexes have been mapped approximately to different regions of

the kinetochore by in vivo measurements with fluorescence microscopy and protein-protein interaction assays (22, 23, 28). Detailed structural analysis is available for a limited number of kinetochore proteins or sub-complexes (24). Inner kinetochore proteins interact with or congregate around centromeric DNA. The DNA-binding interaction of a few proteins has been characterized through deuterium exchange or x-ray crystallography. Investigation of outer kinetochore proteins has focused on how microtubule-binding proteins make direct contacts to the microtubules (22-24). Interestingly, these proteins have been shown to bind microtubules by diverse mechanisms through structural and biochemical analysis(24). Many signaling proteins that regulate or localize to kinetochores have been characterized by x-ray crystallography. These structures have helped to understand regulation of chromosome segregation and cell cycle progression. Based on current data, models of the kinetochore have been proposed that either feature all proteins as cartoon shapes or display the known structures of kinetochore proteins, arranged appropriately between a centromere and a microtubule (22-24). Efforts for more structural analysis are underway to understand the rest of the inner and outer kinetochore proteins, including how the two regions connect.

The nuclear pore complex The nuclear pore complex (NPC) facilitates transport of molecules in and out of the nucleus. As its name and function suggest, the complex resides in the nuclear envelope and is shaped as a squashed sphere with a pore in the middle (29, 30). EM images at increasing resolution give more details, including an 8-fold symmetry, multiple distinct rings, attachments on both sides for flexible filaments, and a basket structure attached to the filaments on the nuclear side. Approximately 30 proteins, called nucleoporins, have been identified through purification and mass spectrometry analysis of this 40-50 MDa complex. The nucleoporins have been shown to organize into sub-complexes and to localize in specific regions of the NPC. Bioinformatic analysis predicted only a few fold types in nucleoporins, including α-helical solenoids and beta-propellers (31, 32). Major ongoing efforts to crystallize individual or pairs of nucleoporins have partially confirmed these predictions, although discrepancies have been observed, particularly in the topology of the α-helical proteins (29). Placement of these crystal structures into EM images cannot yet be accomplished for the entire complex, since the highest resolution EM model is too low (~58 Å) (33). However, an EM model of a subcomplex (the Y- complex) at higher resolution (~35 Å) can be fit tentatively with crystal structures of the

15 constituent proteins (34). As mentioned earlier, the Sali and Rout labs generated a model of the entire nuclear pore complex by integrating multiple sources of data, including comprehensive data on the volumes and the abundance of each nucleoporin (13, 14). Combination of this computational framework with more structural characterization by x-ray crystallography, single- particle cryo-EM on sub-complexes, and other methods will continue progress toward a high- resolution model of the NPC (29, 30).

The spindle pole body The spindle pole body (SPB) is the microtubule-organizing complex (MTOC) in Saccharomyces cerevisiae (35, 36). This large protein complex functions to nucleate and anchor the minus-ends of microtubules. It organizes both nuclear and cytoplasmic microtubules, necessary for chromosome segregation and nuclear positioning, respectively. The SPB also serves as the site of karyogamy (fusion of two yeast nuclei after mating) and as a scaffold for cell-cycle signaling proteins. To understand these functions in detail, a model for the SPB structure is needed. High-resolution structures of the SPB by x-ray or electron crystallography have not been determined because it is difficult to purify the SPB, due to its low copy number in cells, large size, heterogeneous composition, and association with the nuclear membrane. A low-resolution structure of the SPB has been determined by cryo-EM and electron tomography experiments (37- 40), and exhibits a cylinder with multiple layers (Figure 1.1). The layers include the outer plaque (OP), two intermediate layers (IL1+IL2), central plaque (CP), and inner plaque (IP). The inner plaque attaches to nuclear microtubules (NM) and the outer plaque attaches to cytoplasmic microtubules (CM). The central plaque is co-planar with the nuclear envelope (NE) and the half bridge (HB), an electron-dense extension on one side of the core cylinder.

Figure 1.1: Cryo-EM image of SPB (modified and reproduced with permission from Knop, M., Pereira, G., and Schiebel, E. (1999) Biol Cell 91, 291-304. © the Biochemical Society, http://www.biolcell.org ). Abbreviations listed in text.

Approximately 20 SPB protein components have been identified by biochemical and genetic studies and localized to the SPB by microscopy studies (Figure 3.7) (35, 41). Most SPB proteins are essential for cell viability and deletion mutants of non-essential SPB proteins have slow growth and SPB defective phenotypes. The first components were identified by antibodies raised to purified SPB protein components and included Spc42, Spc110, and Spc98 (42). Additional proteins were later identified by mass spectrometry of purified SPB protein components, including Spc29, Spc72, Bbp1, Cnm67, and Nud1 (27). Genetic screens for mutants with SPB defective phenotypes provided Cdc31, Cmd1, Kar1, Mps1, Mps2, Mps3, and Ndc1 (35, 43). The proteins Spc97, Sfi1, and Nbp1 were identified through their interactions with other SPB components in yeast co-affinity purifications or dosage suppressor screens (35, 44). Tub4 was found through genomic analysis to be the yeast homolog of γ-tubulin that is found in MTOC of other organisms (45). Additional detailed knowledge about the SPB proteins is presented in Chapters 2 and 3. Models of SPB have been proposed based on available data (35, 36, 41). The models are usually portrayed as cartoons, describing where each protein resides as shapes within the different regions of the complex. In Chapter 3, I present models of the SPB that incorporate all available structural information for the complex and each individual SPB protein (Figures 3.7 and 3.8). Subsets of SPB proteins have been modeled in more detail. A geometrically-defined model of the central plaque has been devised from in vivo FRET measured distances between central plaque proteins and the hexagonal symmetry observed from Spc42 overexpression (46). A sub-complex that includes Tub4/γ-tubulin has been proposed to interact in ring as a template for microtubule growth, based on single-particle cryo-EM images of the sub-complex (47). Based on crystal structures of Sfi1 and Cdc31 in complex, these proteins have been inferred to assemble into filamentous bundles and to form the site within the half bridge for SPB duplication (48). Duplication of the SPB has been shown to coordinate with the cell cycle to ensure generation of only one SPB per division (38, 49). Through EM studies, the duplication pathway has been mapped out in the following steps: elongation of the HB, assembly of a new SPB at the cytoplasmic tip of the HB, insertion of the SPB into the nuclear envelope, and separation of the two SPB to opposite ends on the nucleus. Despite this description, many questions remain, including how duplication of the SPB is triggered to start, how the proteins are assembled, and

how the size of the SPB is regulated. A detailed structural model of the SPB may help to answer these questions about SPB duplication and help to understand the mechanisms of SPB function. The SPB is often compared to the centrosome, the MTOC of animal cells (35, 50). The centrosome has similar functions and some homologous components as the SPB, but it is much larger (~ 10 x) and morphologically distinct from the SPB, and has additional functions and protein components (50-54). In the center of the complex is a pair of perpendicular centrioles that are symmetric cylinders of 9 triplet microtubules, closely associated with specific proteins. Around the centrioles, a matrix of many other proteins associate in a nebulous sphere called the percentriolar material (PCM). The PCM proteins are responsible for nucleating and organizing microtubules. Despite the amorphous ultrastructure, the PCM is organized into smaller complexes and the proteins are localized to specific different areas within the PCM. The γ- tubulin ring complex, consisting of γ-tubulin and several γ-tubulin binding proteins (GCP2-6,- WD), nucleates and caps the microtubules by forming a ring template for microtubule assembly (55). Pericentrin/kendrin, AKAP450/CG-NAP, ninein, and other proteins anchor the microtubules to different parts of the centrosome. Mutations in centrosomal proteins can lead to defects in centrosomal assembly or regulation, resulting in cancer or various genetic diseases (53, 56). Centrosomal proteins are conserved among animals, but only a few proteins have orthologs in the SPB (35, 54). The orthologous proteins include Tub4 = γ-tubulin, Spc97 = GCP2, Spc98 = GCP3, Spc110 = pericentrin/kendrin, Nud1 = centriolin, Cdc31 = centrin 3, and Sfi1 = hSfi1p. Though the majority of proteins from centrosomes and SPBs are not homologous, one striking similarity is the presence of one or more coiled-coil domains in most of the proteins (~75%) in each complex and few other recognizable motifs or domains (35, 54), suggesting an important common function of coiled coils in the structure of these complexes.

Coiled coils – predicted structural features in the SPB The coiled coil is an interaction domain that forms a structure of two or more intertwined

α-helices (57-60). Coiled-coil sequences have a distinctive heptad pattern [abcdefg]n for every 2 turns of a helix. The left-handed twist between the helices makes up the difference between the 3.6 residue periodicity of helices and the 3.5 residue periodicity of half of a heptad repeat. The a and d residues are often hydrophobic and interact in the core of the twist with “knobs-into-holes” packing. This packing involves a sidechain (knob) from one helix fitting into a space (hole)

between 4 sidechains on another helix. The e and g residues are often charged or hydrophilic, shield the core from solvent, and contribute to interaction specificity in dimers. In higher-order coiled coils, these residues can also contribute to core hydrophobic interactions. The heptad repeat allows identification of coiled coils in protein sequences using computer algorithms with good sensitivity and accuracy (61-63). Many of these programs depend on a training database of known coiled coils, thus programs often have improved performance with an updated, larger, and more diverse training database (64). Coiled coils have been predicted to exist in ~10% of eukaryotic proteins (65), but only a fraction have been experimentally characterized. They can be quite specific in their interactions as homo- or hetero- oligomers, and can vary in their length and topology of the number of helices and helix orientation. Long coiled coils often have disruptions in their heptad repeat, called skips, stutters, and stammers, which can have functional importance (66). Insight into sequence determinants for coiled coils’ specificity and topology have been provided by experimental studies of mutations in particular coiled coils or systematic interaction studies on a group of coiled coils (67, 68). It was found that small changes in sequence or conditions can change the orientation, stoichiometry or partnering of coiled coils, which makes it difficult to predict these features in native sequences. However, some particular residues in heptad positions have been observed to confer specific topologies. For example, an Asn in the a position on two partner helices stabilizes parallel, dimeric coiled coils. Using these types of observations, design of specific coiled coil interactions has been successful for a number of target coiled coils (69). Coiled coils mediate interactions in a wide array of proteins (57, 58). Transcription factors, such as bZIPs, b/HLH/ZIP, and HTH-ZIP families, dimerize via their coiled-coil (ZIP, zipper) domains and bind to different DNA sequences based on the coiled-coil interaction specificity. Proteins involved in vesicular trafficking use coiled coils to aid in membrane fusion (SNAREs), tethering of vesicles (Golgins, p115), and sorting proteins (ESCRTs). Many oncogenes are fusion proteins between a coiled-coil protein and a kinase (ex BCR-Abl), with dimerization of the coiled coil allowing constitutive kinase activity and cancer growth. Coiled coils, especially long ones, have many functions beyond acting as an oligomerization or interaction domain. They frequently serve as stalks, levers, crosslinkers, flexible arms, or springs (66). Intermediate filaments are cytoskeletal stalks composed of

polymerizing coiled-coil proteins, such as keratins and nuclear lamins. Coiled coils in motor proteins have elastic or lever mechanical functions and connect between globular domains bound to microtubules or microfilaments and globular domains bound to cargo. SMC (structural maintenance of chromosomes) proteins have flexible coiled-coil arms that help organize chromosomes. Prior to the work described in this thesis, coiled coils had been detected computationally in SPB protein sequences, but it was not known whether the coiled-coil predictions were correct and if so, how the coiled coils form and function within the SPB. The coiled coils in some of the proteins, such as Spc110 and Cmn67, had been shown to help organize the SPB structure by forming support stalks between different layers (70, 71). Other potential roles of coiled coils in SPB proteins included connecting components through protein-protein interactions, either as homo- or hetero-oligomers, or folding intra-molecularly as part of a protein’s structure. The prevalence and conservation of coiled coils in homologous complexes suggest that coiled coils are important in assembling the SPB. Reducing experimental analysis of the SPB proteins to their coiled coils (or other domains) takes advantage of proteins' modular nature. Well-defined domains can often be isolated from the rest of the protein, folded in isolation into their characteristic structure, and tested experimentally with much greater ease than entire proteins. In particular, coiled coils often fold and interact cooperatively and reversibly, aiding experimental analysis (72, 73). While coiled coils often can be identified computationally, identification of their interaction partners must be determined experimentally. Fortunately, potential interaction partners for coiled-coil helices are other coiled-coil helices. Several efforts to search for partners of native coiled coils have been performed by yeast two-hybrid assays (74, 75). More interactions between proteins can be uncovered in high-throughput assays when limiting analysis to particular domain compared to using whole proteins. Focusing the search to coiled coils from proteins associated with a particular function or complex, as I have done with SPB coiled coils, can further increase the likelihood of finding partners for the coiled coils.

Summary of thesis My thesis work focused on characterizing structural aspects of the SPB, in particular the coiled-coil domains within the constituent proteins. I hypothesized that coiled coils are involved in protein-protein interactions between proteins within the complex. This hypothesis was tested by experimentally determining if, which, and how putative SPB coiled coils interact. Good methods exist to predict the propensity of a sequence to form a coiled coil (62, 64), but it is much more difficult to predict the precise details of the structure, including the orientation, stoichiometry, or partnering preference of the coiled coil. This information is important for defining structural features of the SPB complex. For this work, coiled coils in SPB proteins were predicted with computer algorithms. Twenty-nine coiled-coil domains were cloned in isolation from the rest of the protein. Each coiled coil was expressed, purified, and assayed for self- and hetero-association. I report the identification and characterization of four strongly self-associating coiled coils and several weaker self- and hetero-associating coiled coils from core SPB proteins in Chapter 2. I report the identification and characterization of three strongly self-associating coiled coils and one hetero- associating coiled coil from membrane-associated proteins in Chapter 3. Also in Chapter 3, I combine all known structural information about the SPB and its proteins to assemble a model of its structure. In Chapter 4, I suggest future studies for understanding the role of coiled coils in the SPB and other complexes.

CHAPTER 2

Analysis of coiled-coil interactions between core proteins of the spindle pole body

Reproduced with permission from: Zizlsperger N, Malashkevich VN, Pillay S, Keating AE. (2008) Analysis of coiled-coil interactions between core proteins of the spindle pole body. Biochemistry 47, 11858-68. Copyright 2008 American Chemical Society.

Supporting information: This paper included supplemental tables and figures available via pubs.acs.org. These tables and figures have been incorporated into this thesis chapter and indicated by “S”.

Collaborator notes: Vladimir N. Malashkevich solved the crystal structure of the Spc42 coiled coil. Shirin Pillay cloned, expressed, and purified some of the peptides and tested a few of the peptides by CD and FRET.

ABSTRACT The spindle pole body (SPB) is a multi-protein complex that organizes microtubules in yeast. Due to its large size and association with the nuclear membrane, little is known about its detailed structure. In particular, although many SPB components and some of the interactions between them have been identified, the molecular details of how most of these interactions occur are not known. The prevalence of predicted coiled-coil regions in SPB proteins suggests that some interactions may occur via coiled coils. Here this hypothesis is supported by biochemical characterization of isolated coiled-coil peptides derived from SPB proteins. Formation of four strongly self-associating coiled-coil complexes from Spc29, Spc42, and Spc72 is demonstrated by circular dichroism (CD) spectroscopy and a fluorescence resonance energy transfer (FRET)

assay. Many weaker self- and hetero-associations were also detected by CD, FRET, and/or crosslinking. The thermal stabilities of nine candidate homo-oligomers were assessed; six unfolded cooperatively with melting temperatures ranging from < 11 °C to > 50 °C. Solution studies establish that coiled-coil peptides derived from Spc42 and Spc72 form parallel dimers, and this is confirmed for Spc42 by a high-resolution crystal structure. These data contribute to a growing body of knowledge that will ultimately provide a detailed model of the SPB structure.

INTRODUCTION The spindle pole body (SPB) is the microtubule organizing center of Saccharomyces cerevisiae (35, 36). This large multi-protein complex functions to coordinate and nucleate both nuclear and cytoplasmic microtubules and is necessary for chromosome segregation, nuclear positioning, and karyogamy. Electron microscopy and tomography experiments have provided a low-resolution structure of the SPB (37, 39, 40), showing a multi-layered short cylinder spanning the nuclear membrane, with cytoplasmic and nuclear microtubules connected to the outer and inner layers, respectively. A “half-bridge” structure, located on one side of the SPB and within the nuclear membrane, serves as the site of SPB duplication. Genetic and biochemical studies have identified approximately 20 proteins that comprise the SPB complex (27, 35, 41, 43). Using immuno-EM, proteins have been localized to regions of the SPB (core, membrane-associated, half-bridge) or particular layers or plaques of the SPB core (outer, intermediate 1, intermediate 2, central, inner) (41, 49) (Figure 1). Given the low diversity of proteins comprising the SPB, up to 1000 copies of each are required for assembly of the ~ 300-500 MDa complex (27, 35). Three-dimensional structural information is key to understanding the mechanism and function of important complexes such as the SPB. High-resolution structures are not available for the complex and are known for only three of the individual components: calmodulin (Cmd1) (76), γ-tubulin (Tub4) (77), and centrin (Cdc31) (78). A structure of Cdc31 in complex with part of component Sfi1 has also been reported (48). Interactions between SPB proteins have been determined using co-immunoprecipitation, genetic synthetic lethality, yeast two-hybrid assays, and FRET experiments (27, 35, 42, 46). For some SPB protein-protein interactions, the regions of interaction have been localized by disruption of interaction via point or deletion mutants (35, 41).

Most SPB proteins are predicted to contain one or more coiled-coil domains (27, 35). Coiled coils are a common structural motif, consisting of two or more α-helices wrapped around each other with a left-handed twist (57-59). Coiled-coil sequences exhibit a distinctive heptad pattern, [abcdefg]n, that allows prediction of coiled-coil-forming regions using computer algorithms (61-63). The a and d positions are usually hydrophobic and pack into the coiled-coil interface. Coiled coils can associate intra- or inter-molecularly, as homo- or hetero-oligomers, and in a parallel or antiparallel orientation. The motif often contributes to fiber formation or functions as an oligomerization or interaction domain. Coiled coils are common in cytoskeletal proteins, motor proteins, transcription factors, and membrane fusion proteins (57, 66). Whether coiled-coil structures actually form within or between proteins of the SPB has not been confirmed experimentally. If present, they may serve to connect components through protein-protein interactions, either as homo-oligomers or hetero-oligomers, or serve as intramolecular structural features within individual components. In this study, we computationally identified and experimentally isolated predicted coiled-coil regions of SPB proteins and tested whether and how they interact. Coiled coils often retain their ability to fold when isolated from a larger protein context (72, 73). We focused on the 10 core proteins of the SPB complex in this study. These proteins are the most likely to interact, based on their close proximity within the complex and prior evidence supporting protein-protein interactions among them. We characterized four strongly associating coiled-coil homo-oligomers. Our data also support the possible significance of five weakly associating homo-oligomers and two hetero- oligomers. The high-resolution structure of a coiled-coil homodimer from Spc42 corroborates our solution assays and contributes to atomic-resolution data describing the SPB.

MATERIAL AND METHODS Coiled-coil prediction. Coiled-coil regions were identified in SPB protein sequences using Paircoil2 (64) with a P score cutoff of 0.96 and a 28-residue window (Figure 2.1 and Tables 2.1 and 2.S1). Many of the regions were also high-scoring according to the program Marcoil (62). Predicted coiled-coil regions were manually inspected, and N- and C-terminal ends were chosen based on a sharp drop in score (often at a Pro or Gly residue), a shift in register, and/or a P score of ≤ 0.93 (often corresponding to the loss of the hydrophobic/hydrophilic patterning). The first and last residues were chosen to occupy predicted b, c, or f positions, with the intent of directing

linker sequences away from the coiled-coil interface. Because more than one coiled-coil region was predicted in some proteins, all peptides were named by appending “_#” after the protein name, where # designates the order of the detected coiled coil in the N-to-C terminal direction. Predictions for Cnm67 were ambiguous with respect to whether there may be one long coiled coil (Cnm67_2) or two shorter coiled coils (Cnm67_3 and Cnm67_4) in this protein; constructs accounting for both possibilities were tested (Figure 2.S1). Multicoil was used to predict whether each coiled-coil region was more likely to form a dimer or a trimer (79).

Table 2.1: Predicted coiled-coil regions of core SPB protein sequences. coiled-coil systematic beginning ending Paircoil2 dimer trimer name name residue residue P score probabilitya probabilitya Cnm67_1 YNL225c 159 263 1.00 0.666 0.252 Cnm67_2 YNL225c 303 461 1.00 0.952 0.496 Cnm67_3 YNL225c 303 384 1.00 0.952 0.496 Cnm67_4 YNL225c 369 461 1.00 0.898 0.173 Spc29_1 YPL124w 7 45 0.99 0.477 0.098 Spc29_2 YPL124w 129 174 1.00 0.094 0.246 Spc42_1 YKL042w 53 137 1.00 0.946 0.492 Spc42_2 YKL042w 178 213 0.99 0.002 0.032 Spc42_3 YKL042w 246 315 0.99 0.017 0.068 Spc72_1 YAL047c 97 160 1.00 0.235 0.428 Spc72_2 YAL047c 306 469 1.00 0.268 0.389 Spc72_3 YAL047c 568 622 1.00 0.464 0.228 Spc97 YHR172w 171 234 0.98 0.001 0.032 Spc98_1 YNL126w 622 690 0.97 0.001 0.029 Spc98_2 YNL126w 789 832 0.97 0.000 0.005

a determined using Multicoil (79). The higher probability, indicating the predicted oligomer, is in bold.

Table 2.S1: Predicted coiled-coil regions of core SPB protein sequences. Residue numbers are given in Table 1. Peptides were expressed with the following extra tags and linkers:

N-terminal cysteine: SYKCGGS-coiledcoil-EFGDYKDDDDKGHHHHHH C-terminal cysteine: SYKGS-coiledcoil-EFGGCGGDYKDDDDKGHHHHHH coiled-coil systematic name name coiled-coil sequence KIFLQNSLSKEDFRMLENVILGYQKKVIELGRDNLRQEERAN SLQKELEAATKSNDKTLDNKKKIEEQTVLIENLTKDLSLNKE Cnm67_1 YNL225c MLEKANDTIQTKHTALLSLTDSLRKAELFEI IKGFLAASQQEELSRISQRFKNAKAEAEDLRNELENKKIEIQT MREKNNTLIGTNKTLSKQNKILADKFDKLTIDEKEILKGANE EIKIKLERLNERLGSWEKSKEKYETSLKDKEKMLADAEKKT Cnm67_2 YNL225c NTLSKELDNLRSRFGNLEGNTSERITIKNILQS IKGFLAASQQEELSRISQRFKNAKAEAEDLRNELENKKIEIQT Cnm67_3 YNL225c MREKNNTLIGTNKTLSKQNKILADKFDKLTIDEKEILKG DKFDKLTIDEKEILKGANEEIKIKLERLNERLGSWEKSKEKYE TSLKDKEKMLADAEKKTNTLSKELDNLRSRFGNLEGNTSERI Cnm67_4 YNL225c TIKNILQS Spc29_1 YPL124w GNSASKKFQDDTLNRVRKEHEEALKKLREENFSSNTSEL ASQNVIDDQRLEIKYLERIVYDQGTVIDNLTSRITRLESFILNSI Spc29_2 YPL124w S EEYKRNTEFINKAVQQNKELNFKLREKQNEIFELKKIAETLRS Spc42_1 YKL042w KLEKYVDITKKLEDQNLNLQIKISDLEKKLSDANSTFKEMRF Spc42_2 YKL042w SNTSDQDSRLKAIERTLSVLTNYVMRSEDGNNDRMS SDDDIMMYESAELKRVEEEIEELKRKILVRKKHDLRKLSLNN Spc42_3 YKL042w QLQELQSMMDGDDNIKLDNVSKHNHATH SLGNDTDFRNSIIEGLNLEINKLKQDLKAKEVEYQDTLQFVQ Spc72_1 YAL047c ENLENSESIVNTINHLLSFILT EYDQFINSIRLKFEKSQKLEKIIASKLNEQSHLLDSLELEENSS SVIEKQDHLISQLKEKIESQSVLINNLEKLKEDIIKMKQNEKV LTKELETQTKINKLKENNWDSYINDLEKQINDLQIDKSEEFH VIQNQLDKLDLENYQLKNQLNTLDNQKLILSQYESNFIKFNQ Spc72_2 YAL047c NLL NKELTLRIEELQRRWISERERRKLDANASEARIKALEQENESL Spc72_3 YAL047c RSKLFNLSINNP RELEQIINETEVNKQMELLYNIYEEIFREIEERRTNQSSQEDFN Spc97 YHR172w NFMDSMKNESSLHRLMVAFD PLIRDIINKLSRISILRTQFQQFNSKMESYYLNCIIEENFKEMTR Spc98_1 YNL126w KLQRTENKSQNQFDLIRLNNGTIE KMNLNDHEASNGLLGKFNTNLKEIVSQYKNFKDRLYIFRAD Spc98_2 YNL126w LKN

Figure 2.1: A schematic of the SPB core complex shows the protein composition and the layer(s) in which each component resides (35, 41). Proteins with known structures or structural properties (Tub4, Cmd1, and Spc110) are represented with shapes at the far left. Proteins of unknown structure are indicated with colored lines scaled to indicate protein length. Black bars indicate predicted coiled-coil regions. Membrane-associated and half- bridge proteins are not shown.

Paircoil2 score -10

-20

-30 0 100 200 300 400 500 600 residue

Figure 2.S1: Paircoil2 score for Cnm67. Predicted coiled-coil regions that were cloned are indicated by bars above the graph and specified in Table 2.1.

Domain annotation. Existing computational methods were used to annotate protein domains, folds, and motifs in SPB proteins. The following programs were run using default settings: mGenThreader (80) (only confidence levels of certain or highly probable were considered), DOMAC (81), and InterProScan (82). Protein production. Coding sequences corresponding to the individual coiled-coil regions of each SPB protein were generated using oligonucleotide assembly by PCR, with codons optimized for expression in Escherichia coli (83). Synthetic genes were ligated into a modified pET43a vector (Novagen) that contained coding sequence for a His6 tag for purification, a Flag tag for solubility, a tyrosine for detection by absorbance at 280 nm, and a cysteine either at the N or C terminus for differential labeling (Table S1). The sole native cysteine among all predicted coiled coils (Spc98_1 position 655, predicted coiled-coil heptad position b) was mutated to a serine. Plasmids were transformed into E. coli BB101 or BL21 cells (84). Expression of proteins

was induced with 1 mM IPTG at OD600 = 0.4 - 0.6 for 4 hours. Proteins were purified from cell lysate in 6 M guanidine HCl with NiNTA resin (Qiagen), eluted with 60% acetonitrile and 0.1% trifluoroacetic acid, lyophilized, and stored at 4 °C. Protein concentrations were determined using the method of Edelhoch (85). Protein purity was greater than 90-95% by SDS-PAGE. The molecular weights of the proteins were verified by MALDI-TOF spectrometry to 0.1% accuracy. For crosslinking experiments, synthetic genes were ligated into pSV282 (Vanderbilt University

Medical Center, Center for Structural Biology) containing coding sequences for a His6 tag, maltose binding protein (MBP), and a TEV cleavage site. Plasmids were transformed into E. coli

BL21 cells (84). Expression of proteins was induced with 1 mM IPTG at OD600 = 0.4 - 0.6 for 4 hours. Proteins were purified under native conditions with NiNTA resin (Qiagen), dialyzed into 50 mM sodium phosphate and 150 mM NaCl, pH 7, and stored at -80 °C. For crystallization, S. cerevisiae DNA encoding residues 65-138 from Spc42 was amplified by PCR and ligated into the expression vector pAED4 (86). The resulting plasmids were transformed into E. coli BL21 pLysS cells (84). Expression of proteins was induced with 0.4 mM

IPTG at OD600 = 0.4 for 2 hours. Proteins were purified in 8 M urea on a G-50 sephadex size exclusion column (Pharmacia) and subsequently by reverse-phase HPLC (Vydac C18 column) (87). Unique cysteines were reacted with either iodoacetamide (Sigma), fluorescein 5-maleimide (Invitrogen), or Alexa 568-maleimide (Invitrogen). Lyophilized proteins were resuspended at

100 μM in 50 mM Tris-HCl, 150 mM NaCl, and 5 M guanidine HCl pH 7.5, reduced with 1 mM TCEP, and incubated with 10- to 1000-fold excess conjugating molecule for 2-4 hours. Alkylated proteins were separated from non-alkylated proteins and free iodoacetamide by reverse-phase HPLC (Vydac C18 column). Fluorophore-labeled proteins were desalted to remove free dye. Labeling efficiency was 80-90% based on SDS-PAGE and analytical reverse- phase HPLC analysis. Circular dichroism spectroscopy. Alkylated proteins were analyzed at 15-50 μM in 50 mM sodium phosphate and 150 mM NaCl, pH 7. Circular dichroism (CD) spectra were measured from 195 – 280 nm in 1 nm increments with 2 second averaging time at 25 °C and/or 4 °C, in triplicate, using an Aviv model 400 CD spectrophotometer with a 0.1 cm cell. For proteins with detectable secondary structure, thermal denaturation curves were measured at 1.5-5 μM in 50 mM sodium phosphate and 150 mM NaCl, pH 7 in a 1 cm cell with stirring. Thermal melts were measured at 222 nm from 4 °C to 50-90 °C in 2 °C increments with 60 second equilibrium time and 30 second averaging time, and refolding was measured by cooling to 4 °C at the same rate. Temperatures for 50% thermal denaturation (Tm values) were approximated as the midpoint between fits to linear folded and unfolded baselines. In the absence of a lower baseline, the signal at the lowest temperature was used to approximate the

value of a constant folded baseline, and an upper limit for the Tm is reported. The same Tm values were obtained within 1 °C whether increasing or decreasing temperature, indicating reversible unfolding and adequate equilibration. Fluorescence resonance energy transfer (FRET) assay. Proteins labeled with fluorescein or Alexa 568 were assayed alone or mixed in pairs in triplicate at 5-10 μM in 50 mM Tris-HCl, 150 mM NaCl, and 1 M guanidine HCl pH 7.5. Samples were prepared and mixed at 10-fold higher concentrations in 5 M guanidine HCl and then diluted. Samples were incubated in wells of black 96-well plates for one hour at 25 °C. Fluorescence was measured with a Perkin Elmer Victor3 fluorescence plate reader at two different excitation and emission wavelengths: donor channel – excite 490 nm, emit 535 nm and FRET channel – excite 490 nm, emit 615 nm. FRET signal was corrected and normalized with the following formula (88): FRET = (Fm - Fa - Fd(Dm/Dd))/Dm, where D = donor channel, F = FRET channel, a = acceptor sample alone (Alexa 568-labeled), d = donor sample alone (fluorescein-labeled), and m = mix of donor and acceptor samples, and averaged over 3 independent assays. Fluorescence was also measured after incubating the

samples for an hour at 4 °C and after the addition of 0.4 M trimethylamine N-oxide (TMAO), both at 25 °C and 4 °C. Control coiled coils (cJun, Fra1, Fos, Creb3, and cortexillin) were included in each assay to provide consistency (Figure S2). Analytical ultracentrifugation. Alkylated proteins were dialyzed against reference buffer (50 mM sodium phosphate and 150 mM NaCl, pH 7) and prepared at three concentrations each. Sedimentation equilibrium samples were spun in a Beckman XL-I analytical ultracentrifuge at 4 °C for 10-18 hours (until equilibrated, as assessed by differences between sequential scans) at 20,000, 25,000, and 38,000 rpm. Concentration was monitored by absorbance at 220-230 nm. Data were analyzed globally with the programs Ultrascan (89) and Heteroanalysis (90). Standard deviation of the fit was < 0.02. Sedimentation velocity samples were spun at 20 °C for 12 hours at 42,000 rpm. Concentration was monitored by interference. Data were analyzed globally with the program Sedanal (91). Standard deviation of the fit was < 0.02. Sednterp was used to calculate partial specific volumes from the amino-acid sequences and solvent density from its composition (92). Crosslinking. Alkylated proteins were resuspended at 50 μM in 50 mM sodium phosphate and 150 mM NaCl, pH 7. MBP-fused proteins were used at 25 μM in 50 mM sodium phosphate and 150 mM NaCl, pH 7. Proteins were assayed alone or mixed in pairs, at 40 μM for alkylated proteins or 5 μM for MBP-fused proteins, and incubated 4 hours at 25 °C after 5 min at 50° C. Bis[sulfosuccinimidyl] suberate (BS3) (Pierce) was added to a final concentration of 0.5 mM to the protein samples, incubated for 10 min, and quenched with 100 mM Tris HCl pH 8.0. Samples were analyzed by SDS-PAGE.

Crystallization. Two crystal forms of Spc42_1 (Table 2.3) were obtained at ambient temperature using the hanging-drop vapor diffusion method (87). For crystal form I, 1 μl of a 10 mg/ml protein solution in 10 mM phosphate pH 8.0 was mixed with 1 μl of reservoir solution

containing 100 mM Tris-HCl pH 8.0, 200 mM MgCl2, 25% PEG 4000, and equilibrated against 500 μl of reservoir solution for several weeks. Crystal form II was grown using similar approach, but with reservoir solution containing 100 mM Tris-HCl pH 8.25, 200 mM CoCl2, and 25% PEG 3350. Crystals were taken directly from the crystallization drop and flash-frozen in a stream of cold nitrogen (X-stream cryogenic crystal cooler, Molecular Structure Corporation). Structure determination. Diffraction data were collected at the X4A beamline at Brookhaven National Laboratories, indexed using DENZO and scaled with SCALEPACK (93). Initial phases

were determined by molecular replacement using the program AMoRe (94) and a search molecule constructed by placing two copies of GCN4-p1 head-to-tail to yield a ~60 residue parallel dimer (95). All side chains were truncated to serine. The correct side chains were built into the model using O and COOT as electron density became apparent during refinement with CNS (87, 96-98). Refinement included rounds of simulated annealing, positional refinement, and individual B-factor refinement, as well as model rebuilding using O. Bulk solvent and anisotropic B-factor corrections were applied throughout the refinement. High anisotropy of the diffraction data was modified with ellipsoidal truncation and anisotropic scaling using the Diffraction Anisotropy Server (99). TLS refinement with Refmac (100) was used to improve final statistics (Table 3). We attempted to improve refinement by testing for the possible influence of twinning using the programs DETWIN, ZANUDA, and the UCLA twinning server(94, 101, 102), but the statistics did not improve. The final models were checked for potential model bias using simulated omit maps calculated with CNS. Graphics were made using PyMOL (103) (Figure 5). The structure contains two chains, A and B, corresponding to residues 66-129 and 67-132 of Spc42p, respectively. 119 water molecules are included. The model exhibits good geometry as determined by PROCHECK (104), with bond lengths and angles within the expected ranges. The Spc42_1 structures in the two crystal forms are very similar (RMSD between Ca atoms is 0.9 Å), although structural superposition reveals slightly different overall bending due to slightly different crystal contacts. Structural differences occur mainly on the surfaces of the two crystal variants, whereas the cores are virtually identical. Because data collected from the crystal of form II have higher resolution and the corresponding structure has better refinement statistics, we refer to crystal form II in all further discussions and this is the form deposited in the PDB (2Q6Q.pdb).

RESUTS

Many coiled coils, but few other domains, are predicted in core SPB protein sequences. Although many coiled coils had been predicted in SPB proteins previously (27, 35), we used the updated program Paircoil2 to identify the locations of putative coiled coils within the core SPB protein sequences. Paircoil2 is a version of Paircoil with improved accuracy and sensitivity (61, 64). All but one of the core SPB proteins were predicted to contain one or more

coiled-coil regions at the cutoff used. A total of 15 coiled-coil regions, ranging in length from 32 to 174 amino acids, were selected for experimental testing (Figure 2.1 and Tables 2.1 and 2.S1). Very few domains or structures other than coiled coils were predicted by a variety of domain- and fold-recognition programs. Tub4 was predicted to have a tubulin fold as expected, given that the structure of the human homolog, γ-tubulin, has been solved (77). The only other protein that was confidently annotated with domains and/or structures other than coiled coils was Nud1. The C-terminal half of Nud1 was predicted to adopt a leucine-rich repeat (LRR) structure. LRR domains are made of repeated structural units of alpha helices and beta strands and are involved in protein-protein recognition (105). These domains may be important for assembling the SPB complex structure.

Several core SPB coiled coils self-associate. Among 18 predicted coiled-coil regions in the SPB core proteins, we analyzed 15 as recombinant peptides using CD and a FRET assay. We excluded Spc110 because it was predicted to contain one long coiled coil (~650 residues) that has already been shown to act as a structural support between the inner and central layers and is highly likely to be a parallel homo- oligomer (46, 70). Paircoil2 predicted a coiled coil in each of Tub4 and Cmd1. These regions mapped to surface helices and did not form coiled coils in available crystal structures and thus were not characterized (77, 78). Coiled coils fold and interact cooperatively and exhibit alpha-helical CD spectra, often with approximately equal minima at 208 and 222 nm. Four SPB peptides, Spc29_2, Spc42_1, Spc72_1, and Spc72_3, met this criterion, with helical content ranging from ~ 40-60% at 25 °C as calculated based on the ellipticity at 222 nm (Figure 2.2A and Table 2.2) (106). Approximately 25-35% of each protein sequence consisted of extra amino acids (tags and linkers) not expected to form alpha-helical structure. Melting temperatures (Tms) were determined for the proteins that folded cooperatively: Spc29_2 (52 °C), Spc42_1 (53 °C), and Spc72_3 (39 °C) (Figure 2.2C and Table 2.2). Spc72_1 showed gradual, non-cooperative loss of helical structure with increasing temperature. Self-association of these four coiled coils was also supported by significant FRET signal (Figures 2.3A and 2.S3). The remaining putative coiled coils did not self-associate at 25 °C, as indicated by random-coil-like spectra (data not shown). However, five peptides, Cnm67_1, Cnm67_2,

Cnm67_3, Spc72_2 and Spc97, gave alpha-helical spectra at 4 °C (Figure 2.2B). Cnm67_1,

Cnm67_2, and Cnm67_3 unfolded cooperatively upon heating, with estimated Tms < 11 °C, < 19 °C, and < 13 °C, respectively (Figure 2.2C and Table 2.2). Spc72_2 and Spc97 unfolding was non-cooperative. Self-association of these five coiled coils was not observed in the FRET assay at 25 °C (Figures 2.3A and 2.S3), but under more stabilizing conditions, i.e. at 4 °C and in the presence of TMAO to counteract guanidine HCl, Spc72_2 and Spc97 showed robust FRET signal. Cnm67_1, Cnm67_2, Cnm67_3, and Spc42_2 consistently showed very weak evidence for self-association under these conditions (Figures 2.3, 2.S3, and 2.S4).

Table 2.2: Summary of biophysical analyses of strongly self-associating, weakly self- associating, and weakly hetero-associating coiled coils of the SPB core proteins. helical helix expected MW for observed class coiled coil content T b m orientation 1, 2, or 3 helices c MW b (%) a,b 24900, Spc29_2 44 [59] 52 ºC ambiguous 8289, 16578, 24867 25000e Spc42_1 57 [71] 53 ºC parallel 13245, 26490, 39735 24800 Spc72_1 38 [48] non-cooperative ambiguous 10299, 20598, 30897 15300

strong Spc72_3 61 [79] 39 ºC parallel 9555, 19110, 28665 17400 Cnm67_1 47 [55] < 11 ºC paralleld ND Cnm67_2 66 [75] < 19 ºC paralleld ND Cnm67_3 60 [74] < 13 ºC paralleld ND Spc72_2 28 [32] non-cooperative ambiguous ND

weak Spc97 27 [35] non-cooperative ambiguous ND

Spc72_1-Spc97 ND ND ambiguous ND Spc72_1-Spc72_2 ND ND ambiguous ND hetero

a calculated based on ellipticity at 222 nm at 25 ºC for strong complexes and 4 ºC for weak complexes; values in brackets correspond to % helicity of the SPB-derived portion of the b c peptide, assuming tags and linkers are not helical (106). ND – not determined. weight(s) closest to the observed weight are underlined. d based on modest differences (Figures 3B, S3, and S4). e The two values correspond to the trimer MW that results from fitting a monomer- trimer equilibrium model to the data. The first value is from sedimentation equilibrium and the second value is from sedimentation velocity.

A B

40 40 25 °C 4 °C 30 30 /dmol) /dmol) 2 20 20 2

10 10 (deg*cm (deg*cm -3 0 -3 0

-10 -10 MRE x 10 MRE x 10 -20 -20

-30 -30 195 215 235 255 275 195 215 235 255 275 wavelength (nm) wavelength (nm)

/dmol) -5 2

-10 (deg*cm -3

x 10 -15 222 nm -20 MRE

-25 0 20 40 60 80 100 temperature (°C)

Figure 2.2: (A) CD spectra for strongly self-associating SPB coiled coils at 25 °C. Protein concentrations: Spc29_2 (●) – 33 μM, Spc42_1 (■) – 28 μM, Spc72_1 (▲) – 47 μM, Spc72_3 (♦) - 28 μM. (B) CD spectra for weakly self-associating SPB coiled coils at 4 °C. Protein concentrations: Cnm67_1 (●) – 32 μM, Cnm67_2 (■) – 32 μM, Cnm67_3 (▲) – 23 μM, Spc72_2 (♦) – 15 μM, Spc97 (o) – 15 μM. (C) Thermal denaturation of self-associating SPB coiled coils monitored at 222 nm. Samples were diluted 1:10 from panels A and B in 50 mM sodium phosphate and 150 mM NaCl, pH 7. Symbols are as in panels A and B. MRE indicates mean residue ellipticity.

0.3 0.3

0.25 0.25

0.2 0.2

0.15 0.15 FRET FRET FRET 0.1 0.1 0.05 0.05 0 0

Spc97 Spc72_1 + Spc72_1 + Spc72_1 + Spc97 + Spc29_2 Spc42_1 Spc72_1 Spc72_3 Spc72_2

Cnm67_1 Cnm67_2 Cnm67_3 Spc97 Spc72_2 Spc98_1 Spc98_1

Figure 2.3: FRET assay of pair-wise interactions. (A) FRET signal for mixtures of fluorophore-labeled SPB coiled coils. Alexa-labeled acceptors are in columns and fluorescein-labeled donors are in rows. Representative data shown are for acceptors and donors with labeled C-termini. Data at left were collected after incubating at 25 °C; data at right were collected after incubating at 4 °C with 0.4 M TMAO. Data from all conditions and for acceptors and donors with labeled N-termini are shown separately in Figure S3. (B) Orientation test for self-associating and hetero-associating SPB coiled coils. Representative data shown for strong self-associations were collected after incubating at 4 °C and data shown for weak self-associations and hetero-associations were collected after incubating at 4 °C with 0.4 M TMAO. Data from all conditions are shown separately in Figure S4. Error bars represent the standard deviation of 3 independent assays. For homo-associations, four mixtures were tested for each protein and results are shown from left to right: N-terminal donor + N-terminal acceptor (solid red), C-terminal donor + C-terminal acceptor (dotted red), N-terminal donor + C-terminal acceptor (solid blue), and C-terminal donor + N- terminal acceptor (dotted blue).

Figure 2.3 continued: For hetero-associations, eight mixtures were tested for each protein pair (A + B) and results are shown from left to right: N-terminal donor on A + N-terminal acceptor on B (solid red), C-terminal donor on A + C-terminal acceptor on B (dotted red), N- terminal donor on B + N-terminal acceptor on A (striped red), C-terminal donor on B + C- terminal acceptor on A (checkered red), N-terminal donor on A + C-terminal acceptor on B (solid blue), C-terminal donor on A + N-terminal acceptor on B (dotted blue), N-terminal donor on B + C-terminal acceptor on A (striped blue), C-terminal donor on B + N-terminal acceptor on A (checkered blue). Red bars show combinations of coiled coils that bring fluorophores near each other in a parallel orientation, and blue bars show combinations of coiled coils that bring fluorophores near each other in an antiparallel orientation. Consistent results exhibiting clear orientation bias were obtained for Spc42_1 and Spc72_3, with weaker evidence for Cnm67_1, Cnm67_2 and Cnm67_3. All showed greater signal in the parallel orientation.

parallel orientation

antiparallel orientation

Figure 2.S2: FRET assay for mixtures of fluorophore-labeled control coiled coils. Alexa- labeled acceptors are in columns and fluorescein-labeled donors are in rows. Data were collected after incubating at 25 °C. Expected interactions: Jun-Jun, Fra-Jun, Fos-Jun, Cre- Cre, Cor-Cor, all with stronger signal in the parallel orientation. Abbreviations: Jun – cJun, A)Fra 25 – ºC Fra1 , Cre – Creb3, Cor – Cortexillin.

A) 25 ºC

parallel orientation

antiparallel orientation

B) 4 ºC

parallel orientation

antiparallel orientation

C) 25 ºC with 0.4 M TMAO

parallel orientation

antiparallel orientation

D) 4 ºC with 0.4 M TMAO

parallel orientation

antiparallel orientation

Figure 2.S3: FRET assay for mixtures of fluorophore-labeled SPB coiled coils under

different conditions. Alexa-labeled acceptors are in columns and fluorescein-labeled donors

are in rows.

A) Homo-associations 0.2 0.2 Spc29_2 Spc42_1 0.18 0.18 0.16 0.16 0.14 0.14 0.12 0.12 0.1 0.1 0.08 0.08 0.06 0.06 0.04 0.04 0.02 0.02 0 0 25 °C 4 °C 25 °C TMAO 4 °C TMAO 25 °C 4 °C 25 °C TMAO 4 °C TMAO

0.3 0.2 Spc72_1 Spc72_3 0.18 0.25 0.16

0.2 0.14 0.12 0.15 0.1 0.08 0.1 0.06

0.05 0.04 0.02 0 0 25 °C 4 °C 25 °C TMAO 4 °C TMAO 25 °C 4 °C 25 °C TMAO 4 °C TMAO

0.2 0.2 Cnm67_1 Cnm67_2 0.18 0.18 0.16 0.16 0.14 0.14 0.12 0.12 0.1 0.1 0.08 0.08 0.06 0.06 0.04 0.04 0.02 0.02 0 0 25 °C 4 °C 25 °C TMAO 4 °C TMAO 25 °C 4 °C 25 °C TMAO 4 °C TMAO

0.2 0.2 Cnm67_3 Spc72_2 0.18 0.18 0.16 0.16 0.14 0.14 0.12 0.12 0.1 0.1 0.08 0.08 0.06 0.06 0.04 0.04 0.02 0.02 0 0 25 °C 4 °C 25 °C TMAO 4 °C TMAO 25 °C 4 °C 25 °C TMAO 4 °C TMAO

0.2 Spc97 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 25 °C 4 °C 25 °C TMAO 4 °C TMAO

B) Hetero-associations 0.3 0.2 Spc72_1 + Spc97 Spc72_1 + Spc72_2 0.18 0.25 0.16 0.14 0.2 0.12 0.15 0.1 0.08 0.1 0.06

0.05 0.04 0.02 0 0 25 °C 4 °C 25 °C TMAO 4 °C TMAO 25 °C 4 °C 25 °C TMAO 4 °C TMAO

0.2 0.2 Spc72_1 + Spc98_1 Spc97 + Spc98_1 0.18 0.18 0.16 0.16 0.14 0.14 0.12 0.12 0.1 0.1 0.08 0.08 0.06 0.06 0.04 0.04 0.02 0.02 0 0 25 °C 4 °C 25 °C TMAO 4 °C TMAO 25 °C 4 °C 25 °C TMAO 4 °C TMAO

Figure 2.S4: Orientation test for self-associating and hetero-associating SPB coiled coils under different conditions. Error bars represent the standard deviation of 3 independent assays. A) Homo-associations. Four mixtures were tested for each protein and results are shown from left to right: N-terminal donor + N-terminal acceptor (solid red), C-terminal donor + C-terminal acceptor (dotted red), N-terminal donor + C-terminal acceptor (solid blue), and C-terminal donor + N-terminal acceptor (dotted blue). B) Hetero-associations. Eight mixtures were tested for each protein pair (A + B) and results are shown from left to right: N-terminal donor on A + N-terminal acceptor on B (solid red), C-terminal donor on A + C-terminal acceptor on B (dotted red), N-terminal donor on B + N-terminal acceptor on A (striped red), C-terminal donor on B + C-terminal acceptor on A (checkered red), N-terminal donor on A + C-terminal acceptor on B (solid blue), C-terminal donor on A + N-terminal acceptor on B (dotted blue), N-terminal donor on B + C-terminal acceptor on A (striped blue), C-terminal donor on B + N-terminal acceptor on A (checkered blue). Red bars show combinations of coiled coils that bring fluorophores near each other in a parallel orientation, and blue bars show combinations of coiled coils that bring fluorophores near each other in antiparallel orientation. Consistent results exhibiting clear orientation bias were obtained for Spc42_1 and Spc72_3, with weaker evidence for Cnm67_1, Cnm67_2 and Cnm67_3. All showed greater signal in the parallel orientation.

Several core SPB coiled coils hetero-associate. In addition to mediating association between multiple copies of the same protein, coiled- coil helices could hetero-associate, either inter- or intra-molecularly. All possible pair-wise combinations of predicted core SPB coiled coils were mixed and tested against each other in the FRET assay under a variety of conditions (Figures 2.3A and 2.S3). Because of their location within the SPB, the putative coiled-coil regions of Spc97 and Spc98 were only tested against the Spc72-derived peptides (49) (Figure 2.1). Because each peptide was expressed independently with an N-terminal and C-terminal cysteine, and then labeled separately with fluorescein- maleimide and Alexa 568-maleimide, confidence in the detected associations was gained by assaying multiple combinations of peptide pairs. Fluorescein-labeled Spc29_2, Spc72_1, and Spc72_2 interacted with multiple peptides, but the same interactions were not always observed for the Alexa-labeled constructs. Five pairs were observed to interact consistently, independent of which fluorophore they were labeled with. In decreasing order of FRET signal, they are Spc29_2:Spc72_1, Spc72_1:Spc97, Spc72_1:Spc72_2, Spc72_1:Spc98_1, and Spc97:Spc98_1. We attempted further characterization to confirm the hetero-associations. CD spectra of potentially interacting components were compared before and after mixing, but no change in signal was detected (data not shown). However, at least one of the participants in each candidate hetero-complex also homo-associated, and this observation is consistent with stronger homo- association relative to hetero-association. Using a crosslinking assay, association of Spc72_1

with both Spc97 and Spc72_2 was detected with the bifunctional amine-reactive chemical BS3 (Figure 2.S5). Hetero-association of Spc98_1 with Spc72_1 or Spc97 was not observed in crosslinking experiments (data not shown), but these pairs gave the weakest FRET signal of all putative hetero-complexes, and the crosslinking assay may be less sensitive than the FRET assay. Interaction of Spc29_2 and Spc72_1 was also not detected by crosslinking (Figure 2.S5), despite strong FRET signal. We judged that the FRET signal was likely caused by non-specific interactions of the fluorescein-labeled constructs. Thus, FRET and crosslinking support weak hetero-association of Spc72_1:Spc97 and Spc72_1:Spc72_2 (Table 2.2), and possibly very weak hetero-association of Spc72_1:Spc98_1 and Spc98_1:Spc97. The Spc29_2:Spc72_1 interaction is likely non-specific.

Spc29_2 + Spc72_1 Spc72_1 + Spc97 Spc72_1 + Spc72_2 M29_2+F72_1 M72_1+F29_2 M97+F72_1 M72_1+F97 F72_2+F72_1 M72_1+F72_2 BS3 + + + - - + + + - - + + + - - + + + - - + + + - - + + + - - protein 1 + + - + - + + - + - + + - + - + + - + - + + - + - + + - + - protein 2 + - + - + + - + - + + - + - + + - + - + + - + - + + - + - +

kDa 50 -

31 -

24 -

6 -

Figure 2.S5: Hetero-associations of Spc29_2 + Spc72_1, Spc72_1 + Spc97, and Spc72_1 + Spc72_2 were evaluated by crosslinking. Each of the proteins was tagged with MBP (M) or Flag (F) to differentiate molecular weights on SDS-PAGE and allow detection of hetero- association vs. homo-association. Spc72_2 was only tagged with Flag because its molecular weight is different from Spc72_1. In the gels, protein 1 refers to the MBP-fused protein in each pair, or to F72_2 when testing against F72_1. Proteins were equilibrated before adding BS3 crosslinker and run on gradient SDS-PAGE. Additional bands in the mixed samples (first lane in each sub-section) support hetero-association and are circled on the gel. Arrows at the bottom indicate lanes that support homodimerization of Spc72_1 (black) and homotrimerization of Spc29_2 (white).

Characterization of core SPB coiled-coil interactions. We determined the molecular weights of the strongly self-associating coiled coils using analytical ultracentrifugation. Sedimentation equilibrium data for both Spc42_1 and Spc72_3 fit well to single-species models, with observed molecular weights corresponding to those expected for dimers (Figure 2.4A and Table 2.2). Small deviations from the expected molecular weights have been observed previously with coiled coils (107, 108). Sedimentation equilibrium and sedimentation velocity data for Spc29_2 gave observed molecular weights between those expected for a dimer and a trimer when fit to a single-species model (Figure 2.4B). However, both data sets fit to a monomer-trimer equilibrium model as well as they fit to a single-species model, giving observed molecular weights close to those expected (Table 2.2). Multicoil predicted Spc29_2 to be a trimer (Table 2.1) and crosslinking experiments also support trimerization (Figure 2.S5). Sedimentation velocity data for Spc72_1 gave observed molecular weights between those expected for a monomer and a dimer when fit to a single-species model (Figure 2.4B and Table 2.2). The data fit as well to a monomer-dimer model, but gave unrealistic observed molecular weights corresponding to ¾ of a monomer and 1½ of a monomer. Although Spc72_1 was predicted by Multicoil to be a trimer (Table 2.1), crosslinking experiments showed significant dimer and no trimer (Figure 2.S5). Concentration dependent CD signal ruled out intramolecular interactions as the origin of helical signal (Figure 2.S6). The lack of cooperativity in the thermal melt, evidence of aggregation under certain conditions, and our observation of multiple hetero- associations by FRET and crosslinking assays suggest that Spc72_1 has a propensity towards self- and hetero-assembly. But it may require the rest of the protein and/or complex to form the functionally appropriate associations.

2.5 3 Spc42_1 Spc72_3 2 2.5 2 230 nm 230 nm 1.5 1.5 1 1 Absorbance Absorbance Absorbance Absorbance 0.5 0.5

0 0 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 2 2 2 2 R -R (cm) R -Rmeniscus (cm) meniscus

0.04 0.02

0 0

-0.04 -0.02 0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5 R2-R 2 (cm) R2-R 2 (cm) meniscus meniscus B

0.8 0.7 0.6 0.5

0.4 g(s*) 0.3 0.2 0.1 0 0 1 2 3 4 5 s* (svedbergs)

0.2 Spc29_2 0.4 Spc72_1 0.15 0.35 0.3 0.1 0.25 0.2 0.05 0.15

delta C (fringes) 0.1 delta C (fringes) 0 0.05 0 -0.05 -0.05 6 6.2 6.4 6.6 6.8 7 6 6.2 6.4 6.6 6.8 7 radius (cm) radius (cm)

Figure 2.4: Characterization by analytical ultracentrifugation. (A) Sedimentation equilibrium data for Spc42_1 and Spc72_3. Data for three concentrations and three speeds were fit globally to a single-species model. A representative trace is shown, with residuals from the fit below. Curves expected for molecular weights corresponding to a monomer, dimer or trimer are shown in solid, dotted, and dashed lines, respectively. (B) Sedimentation velocity data for Spc29_2 and Spc72_1. Data collected at three concentrations were analyzed by Sedanal and fit globally to a single-species model. Representative plots of sedimentation coefficient distributions and fitting of time-resolved concentration differences are shown at

50 µM for each protein. Spc29_2 is shown in black and Spc72_1 in gray with ♦ symbols representing observed data, lines representing fits, and residuals plotted around y = 0.

-8

-9 /dmol) 2 -10

-11 (deg*cm -3 -12 x 10

-13 222 nm 222

-14 MRE

-15 1 10 100 1000 concentration (μM) Figure 2.S6: Concentration dependence of Spc72_1 monitored by CD at 222 nm at 25 ºC.

The FRET assay was designed to distinguish the orientation of coiled-coil helices, in cases where this is unique and well-defined. Parallel coiled coils with fluorophores attached at the same termini were expected to generate more FRET signal than coiled coils with fluorophores on opposite ends. The reverse was expected for antiparallel dimers. Large differences in FRET were not expected for antiparallel trimers or tetramers. This assay was validated using parallel coiled-coil dimers of varying lengths (Figure 2.S2). We measured the FRET signals for all possible pair-wise combinations of labeled core SPB peptides (Figures 2.3B, 2.S3, and 2.S4). Self-associating SPB coiled coils Spc42_1 and Spc72_3 consistently exhibited significantly greater FRET when fluorophores were at the same ends of the helices, supporting a parallel orientation (Table 2.2). Cnm67_1, Cnm67_2, and Cnm67_3 also showed a weak preference for parallel orientation. The remaining self-associating SPB coiled coils, and all hetero-associating SPB coiled coils were not consistently observed to give greater FRET with any labeling scheme. This could be because they interact as antiparallel trimers or tetramers, or in a staggered geometry, giving approximately equal FRET in either orientation. It could also be that these assemblies do not have a preference for a single orientation in the absence of the rest of the protein or complex. A third possibility is that the coiled coils were affected differentially by different combinations of fluorophores.

Crystal structure of Spc42_1.

Crystals of Spc42_1 belonged to space group P212121 and contained one Spc42_1 homodimer in the asymmetric unit (87) (Table 2.3). Spc42_1 forms a parallel dimeric coiled coil (Figure 2.5), in agreement with the solution biophysical characterization and the predictions of Paircoil2 and Multicoil. Residues 67-128 of chain A and residues 67-131 of chain B were well defined in the electron density map. The length of the ordered part of Spc42_1 is ~100 Å, and the diameter of the Spc42_1 dimer is ~24 Å at its widest point (Cα-Cα distances are ~14 Å across the dimer). Residues at the core positions show canonical knobs-into-holes packing (109). Fitting Crick parameters for an idealized coiled coil to residues 67-128 of Spc42_1 gave a superhelical radius (R0) of 4.91 Å, an a-position phase angle (φ) of 0.36 radians, and a superhelical frequency

(ω0) of -0.65 radians/amino acid (109, 110) (Figure 5D). The presence of a tyrosine (Tyr 100) in the d position of the fifth complete heptad is uncommon: only ~ 1.7% of d-position residues are Tyr in coiled coils of known structure that are longer than 21 residues (as judged by running the

program SOCKET on recent releases of the PDB) (111). The crystal structure shows that the Cα- Cβ vectors of the tyrosine side chains from each helix point into the interface with perpendicular packing, as expected for d-position residues in parallel dimers (110, 111), but the hydroxyphenyl groups point out of the core towards the bulk solvent. This same geometry is observed for lysine residues (Lys 97) in d positions at the N-terminus of the coiled coil. In both cases, φ deviates locally to a higher value, as expected because the side chains would leave a gap in the core if the backbone did not compensate by bringing the d positions closer (Figure 2.5D). Other deviations

of the Crick parameters include an expected decrease in R0 and φ at Ala 90, an a position.

Table 2.3: Data processing and refinement statistics for Spc42_1.

Crysta l form I II

Space Group P212121 P212121 Cell Dimensions (Å) a = 40.44 a = 42.86 b = 55.89 b = 57.79 c = 59.80 c = 57.17 α=β=γ=90º α=β=γ=90º Resolution (Å) 20.0-2.2 20.0-1.97 a Rmerge 0.067 0.075 I/sigma(I) 14.0 8.2 All atoms/Waters 1209/119 1192/119 Multiplicity 5.0 4.5 Completeness (%) 98.2 94.7 Unique Reflections (working/free) 6938/336 9799/475 b Rcryst/Rfree 0.212/0.344 0.212/0.297 Average B factor (Å2) 31.50 36.77 R.M.S. Standard Deviations From Ideal Values Bond Lengths (Å) 0.019 0.019 Bond Angles (º) 1.77 1.91

a Rmerge = ∑∑j|Ij - |/∑∑||, where Ij is the intensity measurement for reflection j b and is the mean intensity for multiple recorded reflections. Rcryst/Rfree = ∑|Fobs| - |Fcalc|/∑|Fobs|, where the crystallographic and free R factors are calculated using the working and test reflection sets, respectively. The test reflections included 10% of the total reflections that were chosen before refinement of the initial model and were not used during refinement.

B C D

5.30 0.50 Superhelical radius (R0) a-position phase angle (φ) 5.20 0.45 5.10 )

Å 0.40 5.00

radius ( radius 0.35

4.90 (radians) angle

4.80 0.30

4.70 0.25 65 85 105 125 65 85 105 125 residue number residue number

Figure 2.5: Crystal structure of Spc42_1. (A) Stereoview of the representative 2Fo-Fc electron density map near Tyr100 in the d position. Water molecules are shown as red spheres. (B) Side view. (C) Top view. The maximal diameter is approximately 24 Å with protrusion of side chains. (D) Fitted Crick parameters. The line shows the median value and ♦ symbols show values fit using 7-residue windows centered on the indicated residue. Circled residues are Lys 79, Ala 90, and Tyr 100, discussed in the text.

DISCUSSION The structure of the SPB is not known at atomic resolution for numerous reasons. Because there are only 1-2 copies of the complex in each cell, obtaining sufficient sample to study is difficult. Purified complexes are heterogeneous in their dimensions and in the number of associated proteins, sometimes including microtubule-associated proteins (27). Reconstitution of recombinant components has been achieved for some sub-complexes of the SPB, but not yet for the entire complex (112). Even with a homogeneous SPB preparation, the enormous size and low symmetry of some layers would present crystallographic obstacles. Given these challenges, a strategy of breaking the complex into component parts, characterizing these parts individually, and ultimately assembling them into a larger model is an appealing alternative. Over a decade of work has uncovered the identity of the protein components of the SPB and their approximate locations within the complex (27, 35, 41, 43). The structures of a few individual proteins or domains have been determined (76-78), and we now contribute the structure of a coiled-coil region of Spc42. Interactions between SPB components have been mapped at low resolution using yeast two-hybrid or immunoprecipitation approaches (35, 41, 42, 46). A recent study from the Davis lab devised a model of the central plaque based on in vivo FRET data (46). This model gave improved resolution because fluorescent probes were attached to either end of SPB proteins. Now we further refine aspects of the structure by demonstrating how some of the SPB proteins may interact – via coiled coils. We found many self-associations and few pair-wise hetero-associations between coiled-coil regions of the core SPB proteins. These observations and subsequent biophysical characterization add to the constraints describing the molecular architecture of the SPB, as discussed below. The large size of the SPB requires extensive oligomerization involving multiple copies of each protein. Thus, it is not surprising that the coiled coil, a very common oligomerization domain, is ubiquitous among proteins of the SPB. Coiled coils could play a role in SPB assembly by creating homo-oligomeric modules, by mediating interaction within a multi-protein module, or by linking distinct autonomous modules together. Our data suggest all types of roles. For example, strongly self-associating coiled coils in Spc29, Spc42, and Spc72 likely comprise subunits that may then assemble, via weaker self- or hetero-associating coiled coils or other interactions, into the larger complex. The weak interaction that we observed between Spc97 and Spc98_1 coiled coils may help assemble the γ-tubulin complex, comprised of Tub4, Spc97, and

Spc98. Attachment of the γ-tubulin complex to the rest of the outer plaque may involve the association of the Spc72_1 and Spc97 coiled coils. It is reasonable to assume that very weak interactions observed between isolated domains in vitro may nevertheless be functionally significant in the SPB. Local concentrations in the complex are very high, and avidity may play a large role. Electron micrographs revealed that the central plaque of the SPB consists of a hexagonal lattice (37). Spc42 was sufficient to establish the framework of this lattice, based on its localization to the central plaque and the observation that its over-expression led to the formation of a lattice that extends outside the diameter of the SPB. Electron-dense features in the micrographs of Spc42 lie on a three-fold axis, are ~25 Å in diameter, and have been interpreted as trimers or trimers of dimers (37). Our data supports the latter model, also favored by Muller et al. (46). The trimer of dimers structure from HIV gp41 is ~35 Å in diameter (113), compatible with the low-resolution EM measurement. The available evidence thus leads to a model in which a homodimer formed via the first coiled-coil domain of Spc42 may further associate via a trimeric interface involving another part of the protein. This homo-trimerization is probably not mediated by a coiled coil, given our observations that Spc42_2 and Spc42_3 do not self- or hetero-associate. Spc42 further associates with other proteins, potentially via coiled coils involving Spc42_2 and Spc42_3, to assemble the rest of the SPB complex. The entire coiled-coil region of Cnm67 has been proposed to form a coiled-coil support separating intermediate layer 2 and the outer plaque, similar to the way the coiled-coil region of Spc110 separates the central and inner plaques (70, 71). Paircoil2 did not predict a continuous coiled coil, but rather split the region into 2 or 3 individual coiled coils (Figure 2.S1 and Tables 2.1 and 2.S1). A dramatic dip in score for over 25 residues, including helix-breaking prolines and a glycine and loss of the hydrophobic/hydrophilic patterning, led us to separate Cnm67 into three shorter sequences and test these in four peptides. However, our data are consistent with a parallel coiled-coil homo-oligomer that includes all of the regions tested serving a role as a spacer. We found that each of Cnm67_1, Cnm67_2, and Cnm67_3 form weak self-associating complexes, all of which showed evidence of parallel orientation (Figures 2.2 and 2.3). Multicoil also predicted a strong dimer preference for all three sequences (Table 2.1). This complex may be more stable in the context of the full-length Cnm67 protein or the SPB complex. The coiled-coil

disrupting features detected by Paircoil2 could potentially have structural or functional consequences. Although our data add to knowledge of the structure of the SPB, it is intriguing to speculate about the rest of the complex. Many of the predicted coiled-coil regions were not found to self- or hetero-associate. These sequences may not encode coiled coils, but that cannot be concluded from our data so far. Some coiled coils may require other parts of the protein to form. Weak coiled-coil interactions may not be detectable in our assays, but could still be important in the high-density environment of the SPB. Further, only self- and binary hetero-associations were tested here. Any coiled coils formed from 3 or more components would not be detected. The reagents that we have prepared are well suited for identifying a role for higher-order coiled-coil assembly, and we are developing assays for that purpose. More work is also necessary to determine structures for non-coiled-coil regions, and to address how these assemble into the SPB complex, although this is complicated by the lack of other predicted autonomous domains. As additional data become available, characterized components can be incorporated into models describing the structural details of the SPB. The approach of assembling a model of the whole complex based on constraints derived from multiple data sources was pioneered for the SPB by Muller et al. (46). A similar strategy has recently been used on a larger scale by Sali and colleagues to construct a hybrid computational/experimental model for the nuclear pore complex (13). For the future, an approach of experimental dissection coupled with geometry-guided reassembly is a promising way of advancing our understanding of large protein complexes.

CHAPTER 3

Specific coiled-coil interactions contribute to a global model of the structure of the spindle pole body

A modified version of this chapter is currently in press with Journal of Structural Biology.

ABSTRACT As the microtubule-organizing center of yeast, the spindle pole body (SPB) is essential for cell viability. Structural studies of the SPB are limited by its low copy number in the cell, its large size and heterogeneous composition, and its association with the nuclear membrane. However, low-resolution or indirect structural information about the SPB may be deciphered through a variety of techniques. Interestingly, a large proportion of SPB proteins are predicted to contain one or more coiled coils, a common protein interaction motif. The high frequency of coiled coils suggests that this structure is important for establishing the overall architecture of the complex. Support for this hypothesis was reported previously for coiled coils from some SPB proteins. Here, we extend this approach of isolating and characterizing additional SPB coiled coils, to improve our understanding of SPB structure and organization. Self-associating coiled coils from Bbp1, Mps2, and Nbp1 were observed to form stable parallel homodimers in solution. Coiled-coil peptides from Bbp1 and Mps2 were also observed to hetero-associate. Experimental coiled-coil interaction data from this work and previous studies, as well as predicted and experimental structures for other SPB protein fragments and domains, were integrated to generate a model of the SPB structure.

INTRODUCTION Very few proteins work alone. More often, they work in complexes or pathways that range in size from a few molecules to hundreds or thousands. Exploring the structures of such complexes is key to understanding their functions as well as their mechanisms of assembly and interactions with other cellular components. However, structure elucidation using X-ray crystallography can be challenging for large assemblies. Large quantities of homogeneous

complexes need to be purified, and the preparations must crystallize and diffract well. When direct experimental structure determination is intractable, an effective alternative is to integrate data from many diverse sources into a structural model (1). This type of approach has been used recently to describe the proteasome (20), the mammalian ribosome (114), and the nuclear pore complex (13). Integrative models are less precise than full atomic-level descriptions of structure obtained from crystallography, but can nevertheless provide significant insights and guide future studies. A protein complex for which an integrative approach could be valuable is the spindle pole body (SPB). This ~ 0.5 GDa complex is the microtubule-organizing center of yeast (35, 36). It functions to nucleate cytoplasmic and nuclear microtubules and serves as an anchor at the minus-end of microtubules. High-resolution structures of the SPB by X-ray or electron crystallography have not been possible, due its low copy number in cells, large size, heterogeneous composition, and association with the nuclear membrane, making purification difficult and crystallization improbable. However, a growing body of knowledge does describe the overall shape, components, and assembly of the SPB. The size and shape of the SPB have been observed by electron microscopy (EM), displaying a ~150 x 150 nm core cylinder (Figure 1.1) (37-40). The cylinder has multiple layers, including outer, central, and inner plaques. The outer and inner plaques are attached to cytoplasmic and nuclear microtubules, respectively. The core cylinder is situated within the nuclear envelope that is co-planar with the central plaque. Despite this localization, proteins of the central plaque do not have sequence features indicating transmembrane domains (35). However, other proteins that localize to the periphery of the SPB near the nuclear envelope do contain transmembrane domains (44, 115-119). These membrane-associated proteins are hypothesized to anchor the complex into the nuclear envelope and help insert a newly synthesized complex. In addition to the core cylinder and membrane-associated region, an extension of the complex into the nuclear envelope has been observed on one side of the SPB and is called the half bridge (38). It is the site on which a new SPB complex forms, and it consists of a distinct set of proteins from the core layers (120-125). When the newly formed SPB is large enough, it inserts into the membrane and for a brief time, a full bridge is observed before the two SPBs separate and migrate, each with its own half bridge, to opposite poles of the nucleus (38).

Biochemical purifications of the SPB and genetic screens perturbing the complex have identified the constituent proteins, and the approximate locations of these proteins within the complex have been determined using EM (27, 35, 43, 49, 126). Interactions between SPB proteins have been tested in some combinations (35, 46, 127-129). High-resolution structures are also available for some individual proteins, parts of proteins, and pairs of proteins (48, 76-78, 129). Several groups have combined this information to generate structural models at low resolution for all or part of the complex (35, 36, 46-48). In particular, the Davis group derived a topological model of the central layers that was based on cryo-EM and in vivo FRET data (46). A notable feature of the SPB is that its constituent proteins are enriched with sequences that are predicted to form α-helical coiled coils (27, 35, 129). The coiled coil is a structural motif consisting of two or more intertwined α-helices (57-60). Coiled-coil structure is encoded by a repetitive sequence that can be computationally detected, due to the conservation of hydrophobic and hydrophilic residues at specific sites. Coiled coils often mediate oligomerization or protein- protein interactions. Given the reported prevalence of coiled coils in SPB proteins, we isolated the predicted coiled-coil domains in order to experimentally verify formation of coiled coils and identify specific areas of interaction (129). Motivating this approach is the observation that coiled coils often fold autonomously, although it is also true that some coiled coils do not fold independently outside the context of the native proteins (72-75). We recently reported characterization of coiled coils within SPB core proteins and now we add to this an analysis of coiled coils from membrane-associated and half-bridge proteins. In addition, we predict other structural features of SPB proteins using computational fold recognition algorithms (3, 4, 80, 82). We incorporate our data and predictions with a review of prior structural knowledge to give a working structural model of the SPB complex. This model will help guide future studies.

MATERIALS AND METHODS Coiled-coil prediction and production. Coiled-coil regions in membrane-associated and half- bridge SPB proteins were identified, cloned, expressed, and purified as described previously (Table 3.1) (129). For circular dichroism experiments to confirm the detected hetero-association, synthetic genes were cut to the coding sequence corresponding to the most confident coiled-coil region (residues 236-267 in Bbp1 and 239-270 in Mps2) and ligated into pSV282 (Vanderbilt University Medical

Center, Center for Structural Biology) containing coding sequences for a His tag, maltose 6 binding protein (MBP), and a TEV cleavage site. Plasmids were transformed into E. coli BL21 cells (84). Expression of proteins was induced with 1 mM IPTG at OD = 0.4 - 0.6 for 4 hours 600 at 37 ºC. Proteins were purified under native conditions with NiNTA resin (Qiagen). MBP was removed from the coiled-coil peptide through cleavage with TEV protease. Peptides were purified from MBP and TEV with NiNTA resin (Qiagen) and reverse-phase HPLC (Vydac C18 column). The molecular weights of the peptides were verified by MALDI-TOF spectrometry to 0.1% accuracy. Circular dichroism spectroscopy, analytical ultracentrifugation, and crosslinking experiments. Peptide samples were analyzed as described previously (129). Fluorescence resonance energy transfer (FRET) assay. Proteins labeled with Alexa 488 or Alexa 568 were assayed alone or mixed in pairs in triplicate at 5 μM in 50 mM Tris-HCl, 150 mM NaCl, and 1 M guanidine HCl pH 7.5. For selected peptides or peptide pairs, a mixture of alkylated peptides were mixed with the fluorophore-labeled peptides at equal molar concentration. Samples were prepared and mixed at 10-fold higher concentrations in 5 M guanidine HCl and then diluted. Samples were incubated in wells of black 96-well plates for one hour at 25 °C. Fluorescence was measured with a Molecular Devices SpectraMax M5 fluorescence plate reader at two different excitation and emission wavelengths: donor channel – excite 490 nm, emit 520 nm and FRET channel – excite 490 nm, emit 610 nm. FRET signal was corrected and normalized with the following formula (88): FRET = (Fm - Fa - Fd(Dm/Dd))/Dm, where D = donor channel, F = FRET channel, a = acceptor sample alone (Alexa 568-labeled), d = donor sample alone (Alexa 488-labeled), and m = mix of donor and acceptor samples, and averaged over 3 independent assays. Fluorescence was also measured after incubating the samples for an hour at 4 °C and after the addition of 0.4 M trimethylamine N-oxide (TMAO) at 4 °C. Fluorescence anisotropy (FA) assay. Samples were prepared as described in the FRET assay. After all FRET measurements were taken, the wells with only one fluorophore-labeled peptide present were diluted 1:10 with 50 mM Tris-HCl and 150 mM NaCl pH 7.5 and measured for fluorescence anisotropy with the same plate reader at two different excitation and emission wavelengths, depending on the fluorophore in the well: excite 485 nm, emit 525 nm (Alexa-488 labeled samples) and excite 575 nm, emit 610 nm (Alexa-568 labeled samples). FA was calculated by the following formula: FA = (Ipara – Iperp)/( Ipara + 2* Iperp), where Ipara = intensity

from parallel emission and Iperp = intensity from perpendicular emission, and averaged over 3 independent assays. Domain annotation. Published computational methods were used to annotate protein domains, folds, and motifs in SPB proteins. The following programs were run using default settings: mGenThreader (80) (only confidence levels of certain or highly probable were considered), InterProScan (82), Phyre (4), BioInfoBank Meta Server (3), and HHpred (130).

RESULTS

Many coiled coils are predicted in membrane-associated and half-bridge SPB proteins. In prior work, we characterized putative coiled coils from SPB core proteins (129). We have now extended our analysis to putative coiled coils from membrane-associated (MA) and half-bridge (HB) proteins. The coiled-coils regions were identified within the MA and HB protein sequences using the program Paircoil2 with a P score cutoff of 0.97 (64). Certain coiled- coil regions in Bbp1 and Mps3 contain a dip in the P score, resulting in uncertainty about whether the region forms one long coiled coil (here named Bbp1_1; Mps3_1) or two shorter coiled coils (here named Bbp1_2 and Bbp1_3; Mps3_2 and Mps3_3); thus peptides corresponding to both possibilities were tested. We did not test the predicted coiled coil from Sfi1 because it maps to a consensus α-helical repeat region that binds Cdc31 (48). In fact, at a more lenient prediction cutoff, low-confidence coiled-coil-like regions are identified in a regular pattern throughout the region of Sfi1 where Cdc31-binding is reported. Thus, the one region with a high score is likely a false-positive prediction error. We analyzed 14 predicted coiled-coil regions in the 4 MA and 4 HB proteins as recombinant peptides using circular dichroism (CD) and a FRET assay (Table 3.1).

Many coiled coils from SPB membrane-associated proteins self-associate. Coiled coils fold and interact cooperatively and exhibit α-helical CD spectra, often with approximately equal minima at 208 and 222 nm (131). Four membrane-associated SPB peptides, Bbp1_1, Bbp1_3, Mps2_1, and Nbp1, were found to self-associate, based on this criterion. The helical content of these peptides ranged from ~ 40-60% at 25 °C as calculated based on the ellipticity at 222 nm (Figure 3.1A and Table 3.1). Approximately 25-35% of each peptide sequence consisted of extra amino acids (tags and linkers) not expected to form α-helical

57 structure. The peptides unfolded cooperatively upon heating and stability was assessed by their melting temperatures (Tms): Bbp1_1 (39 °C), Bbp1_3 (41 °C), Mps2_1 (34 °C), and Nbp1 (43 °C) (Figure 3.1C). Self-association of these four coiled coils was also supported by strong FRET signal (Figure 3.2A). The remaining putative coiled coils did not self-associate at 25 °C, as indicated by random-coil-like spectra. However, two peptides, Bbp1_2 and Mps2_2, gave α-helical spectra at

4 °C (Figure 3.1A). They both unfolded cooperatively upon heating, with estimated Tms < 10 °C and < 19 °C, respectively. Self-association of these coiled coils was not observed in the FRET assay at 25 °C, but under more stabilizing conditions, i.e. at 4 °C and in the presence of TMAO to counteract guanidine HCl, Mps2_2 showed evidence of self-association (Figure 3.2A). None of the putative coiled coils from HB proteins self-associated by either assay (Figures 3.1B and 3.2B)

Table 3.1: Predicted coiled-coil regions of membrane-associated and half-bridge SPB protein sequences and summary of CD experiments.

systematic name and helical content protein T (°C) b residues (%)a m Bbp1_1 YPL255W 214-385 40.5 [45.8] 39

Bbp1_2 YPL255W 214-291 42.7 [53.1]* < 10 Bbp1_3 YPL255W 285-385 39.9 [48.2] 41 Mps2_1 YGL075C 151-228 47.0 [58.4] 34 Mps2_2 YGL075C 235-277 35.7 [49.8] < 19 associated membrane - Nbp1 YLR457C 166-222 58.0 [75.7]* 43 Ndc1 YML031W 324-359 random coil* Kar1_1 YNL188W 57-84 random coil

Kar1_2 YNL188W 120-147 random coil Kar1_3 YNL188W 344-371 random coil

bridge Mps3_1 YJL019W 193-283 random coil* - Mps3_2 YJL019W 193-234 random coil half Mps3_3 YJL019W 225-283 random coil Mps3_4 YJL019W 350-393 random coil a All peptides were tested in PBS, except starred peptides were tested in PBS + 0.5 M GuHCl for solubilization. Helical content calculated based on ellipticity at 222 nm at 25 ºC for strong complexes (in bold) and 4 ºC for weak complexes; values in brackets correspond to % helicity of the SPB-derived portion of the peptide, assuming tags and b linkers are not helical. Tm values were approximated as the midpoint between fits to linear folded and unfolded baselines. In the absence of a lower baseline, the signal at the lowest temperature was used to approximate the value of a constant folded baseline, and an upper limit for the Tm is reported.

Figure 3.1: (A) CD spectra for coiled coils from membrane-associated SPB proteins. Data were collected at 25 ºC for strong complexes (in bold) and 4 ºC for weak or non-interacting complexes. (B) CD spectra for coiled coils from half-bridge SPB proteins at 4 ºC. (C) Thermal denaturation of self-associating SPB coiled coils monitored at 222 nm. MRE indicates mean residue ellipticity. All peptides were tested in PBS, except starred peptides were tested in PBS + 0.5 M GuHCl for solubilization.

A B membrane- orphan associated core

Bbp1_1 0 0 0 0 0 0 0

Bbp1_3 0 0 0 0 0 0 0

Mps2_1 0 0 0 0 0 0 0 half-bridge 0 0.1

Nbp1_3 0 0 0 0 0 0 0 0 0.07 Kar1_1 0 0 0 0 0 0 0

Bbp1_2 0 0 0 0 0 0 0 0 0 0 0 0.05 Kar1_2 0 0 0 0 0 0 0

Mps2_2 0 0 0 0 0 0 0 0 0 0 0 0.03 Kar1_3 0 0 0 0 0 0 0

Ndc1_1 0 0 0 0 0 0 0 0 0 0 0 0.02 Mps3_1 0 0 0 0 0 0 0

Spc29_1 0 0 0 0 0 0 0 0.01 Mps3_2 0 0 0 0 0 0 0

Spc42_2 0 0 0 0 0 0 0 0 Mps3_3 0 0 0 0 0 0 0

Spc42_3 0 0 0 0 0 0 Mps3_4 0 0 0 0 0 0 0 Kar1_1 Kar1_2 Kar1_3 Mps3_1 Mps3_2 Mps3_3 Mps3_4 Bbp1_1 Bbp1_3 Mps2_1 Nbp1_3 Bbp1_2 Mps2_2 Ndc1_1 Spc29_1 Spc42_2 Spc42_3 C

Figure 3.2: FRET assay of pairwise interactions. FRET signal for mixtures of fluorophore- labeled membrane-associated and orphan core coiled coils (A) and half-bridge coiled coils (B). Alexa-labeled acceptors are in columns and fluorescein-labeled donors are in rows. Representative data shown are for acceptors and donors with labeled C-termini. Data were collected after incubating at 4 °C with 0.4 M TMAO. FRET signal correlates to color bar on the right. (C) Orientation test for self-associating and hetero-associating SPB coiled coils. Representative data shown were collected after incubating at 4 °C with 0.4 M TMAO. Error bars represent the standard deviation of 3 independent assays. For color schemes, see Figure 2.3B. Red bars show combinations of coiled coils that bring fluorophores near each other in a parallel orientation, and blue bars show combinations of coiled coils that bring fluorophores near each other in an antiparallel orientation. Consistent results exhibiting clear orientation bias were obtained for the self-associations, with weaker evidence for the hetero-association. All showed greater signal in the parallel orientation.

Characterization of homo-oligomerizing coiled coils in membrane-associated proteins. We determined the stoichiometry of the strongly self-associating coiled coils using analytical ultracentrifugation (Figure 3.3 and Table 3.2). Sedimentation velocity data for Bbp1_3 and Mps2_1 fit well to a single-species model with observed molecular weights corresponding to those expected for dimers. Data for Nbp1_3 did not fit well to a single-species or any monomer- N-mer models; however they did fit well to a single-species model with non-ideality included, likely needed to account for the influence of the 0.5 M guanidine hydrochloride in the solvent. The observed molecular weight corresponded to that expected for a dimer. In conclusion, all three coiled coils include two peptide chains. We determined the orientation of the strongly self-associating coiled coils using a FRET assay. In this assay, parallel coiled-coil dimers of peptides with approximately equal length are expected to generate more FRET signal when acceptor and donor fluorophores are attached at the same termini than when they are attached at opposite termini. Conversely, anti-parallel coiled coils are expected to show the reverse behavior. Because each peptide was expressed independently with an N-terminal and C-terminal cysteine, and then labeled separately with donor and acceptor fluorophores, confidence in the detected orientation was gained by assaying multiple combinations of peptide pairs. All three coiled coils consistently exhibited significantly greater FRET when fluorophores were at the same ends of the helices, supporting a parallel orientation (Figure 3.2C).

Table 3.2: Oligomerization of strongly self-associating coiled coils from SPB membrane-associated proteins.

expected MW (Da) for protein 1, 2, or 3 helicesa observed MW (Da) b S c st.dev. Bbp1_3 11228, 22456, 33684 21760 1.6 0.0052 Mps2_1 9177, 18354, 27531 18020 1.4 0.0058 Nbp1_3 6865, 13730, 20595 13680 1.5 0.0051

a Expected weights closest to the observed weights are in bold. bAssessed by c analytical ultracentrifugation. Sedimentation coefficient.

Figure 3.3: Oligomerization of strongly self-associating coiled coils from SPB membrane- associated proteins, assessed by analytical ultracentrifugation. Sedimentation velocity data collected at three concentrations were analyzed by Sedanal and fit globally to a single-species model (with non-ideality for Nbp1_3). Representative plots of sedimentation coefficient distributions and fitting of time-resolved concentration differences are shown at 100 μM for Bbp1_3 (black) and Mps2_1 (dark gray) and 50 μM for Nbp1_3 (light gray). In each graph showing fits to the data, symbols represent observed data, lines represent fits, and residuals are plotted around y = 0.

Coiled coils from Bbp1 and Mps2 hetero-associate. Coiled coils have the potential to self associate, which is easy to assay by CD, but they can also hetero-associate with other coiled-coil peptide chains, either intra- or inter-molecularly. To identify possible two-component hetero-complexes, all possible pair-wise combinations of predicted coiled coils within each SPB subcomplex (MA or HB) were mixed and tested against each other in the FRET assay under a variety of conditions. In addition, coiled coils from core SPB proteins that had no known partner were tested against three MA coiled coils. Bbp1_2 and Mps2_2 were consistently observed to interact in this assay (Figure 3.2A). We confirmed the hetero-association between Bbp1_2 and Mps2_2 with two methods that also provide some information about the likely stoichiometry of the complex. Characterization of the hetero-complex was complicated by the fact that both coiled coils self- associate weakly, although the FRET assay results indicated that the hetero-association was stronger. Under some circumstances, hetero-associating coiled coils can be detected by an increase in helical CD signal in a mixture of two components compared to the average of each of the components measured alone. But if both individual components are highly helical under the assay conditions, little or no change may be observed; this was the case for Bbp1_2 and Mps2_2. We constructed tagless, truncated peptides (denoted by names that begin with ‘t’) that were more soluble, and also resulted in loss of helicity for tBbp1_2 that facilitated heterocomplex detection. At a 1:1 ratio of tBbp1_2 and tMps2_2, we did not observe an increase in signal, but with decreasing amounts of tBbp1_2, an increase in helical content of the mixture relative to the components was observed (Figure 3.4A). This increase in signal reached a plateau at a 1:4 ratio (tBbp1_2 - 10 µM and tMps2_2 - 40 µM), consistent with expectation for a 1:4 complex stoichiometry. The hetero-association and stoichiometry results were supported by a crosslinking assay carried out at a range of peptide concentrations (Figure 3.4B). At the highest concentration, the prominent species was ~45 kDa, which corresponds to the molecular weight expected for either two molecules of Bbp1_2 with two or three molecules of Mps2_2 (41 or 50 kDa), three molecules of Bbp1_2 with 1 molecule of Mps2_2 (46 kDa), or one molecule of Bbp1_2 with four molecules of Mps2_2 (45 kDa). The latter stoichiometry agrees better with the CD results and with the observation that excess Bbp1_2 (as monomer and dimer bands), but not Mps2_2, is present in the mixed, crosslinked 100 µM sample.

-6

nm -8 222 3

- -10

10 -12 -14 MRE x -16 4:1 2:1 1:1 1:2 1:3 1:4 1:5 1:6 ratio Bbp1_2:Mps2_2

B [peptide] 100 µM 30 µM 10 µM BS3 + + + + + + + + + - - - Bbp1_2 + + - + + - + + - + + - Mps2_2 + - + + - + + - + + - + kDa 200 116.3 97.4 66.3 55.4

36.5 31.0

21.5

14.4 6.0

Figure 3.4: Confirmation of hetero-association between Bbp1_2 and Mps2_2. (A) CD spectra for Bbp1_2 and Mps2_2 separately and mixed together at indicated

concentration ratios at 4 °C. The purple triangles of the 1:1 mix are the same in both graphs. Inset graph shows MRE at 222 nm, showing a plateau at 1:4 ratio. (B) Alkylated proteins Bbp1_2 and Mps2_2 were equilibrated at concentrations indicated, with

constant moles and variable volumes, before adding BS3 crosslinker. Samples were concentrated before running on gradient SDS-PAGE. An arrow indicates a band of ~45 kDa that supports hetero-association.

Putative coiled coils were not observed to participate in multi-component interactions. We tested all of the identified putative coiled coils for self-association and binary hetero- association with numerous candidate SPB coiled coils. However, coiled coils may require more than two components to form. For example, coiled coils may interact as hetero-trimers or hetero- tetramers. We tested for possible multi-component interactions in two ways. First, we modified the FRET assay such that each pair of fluorophore-labeled peptides was incubated with a mixture of unlabeled peptides. An increase in FRET signal under these conditions would indicate that a component of the unlabeled peptide mixture participates in a multi-component complex with the labeled peptide pair. Second, we monitored the fluorescence anisotropy (FA) for each individual fluorophore-labeled peptide alone and incubated with the mixture of unlabeled peptides. An increase in signal would indicate interaction between one or more of the unlabeled peptides with the fluorophore-labeled peptide. We focused the multi-component studies on peptides that did not self-associate or hetero- associate with an identified partner, i.e. the coiled-coil “orphans”. Of the 12 putative coiled coils from core SPB proteins, four were found to self-associate strongly, four were found to self- associate weakly, and we believe it is likely that all of the separate regions in Cnm67 likely form one longer coiled coil (possibly with disruptions) (129). That leaves 3 core coiled coils as orphans: Spc29_1, Spc42_2, and Spc42_3. Of the 7 putative coiled coils from membrane- associated SPB proteins, four were found to self-associate and two were found to hetero- associate, leaving one orphan coiled coil: Ndc1. There are many reasons that our methods may not have identified interaction partners for these peptides: The predicted coiled-coil regions may be false positives from the coiled-coil prediction algorithm; these peptides may not form coiled coils in isolation from the rest of the protein; the interactions of these peptides may be too weak to be detected by our assays out of the context of the SPB; or they may only form coiled coils in hetero-associations with more than two partners. We addressed the last possibility by testing orphan coiled coils against each other, including in the presence of mixes containing all core or all membrane-associated coiled coils. By both the FRET and FA assays, no multi-component interactions were identified for the orphan coiled coils (Figures 3.5 and 3.6A). Confidence that we can detect at least strong interactions in these assays derives from two indirect controls: strongly self-associating core coiled coils, Spc29_2 and Spc72_3 (Figures 3.5 and 3.6A). In the FRET assay, signal for these coiled coils decreased when samples were

mixed with unlabeled core coiled coils relative to fluorophore-labeled pairs alone, but stayed the same when mixed with unlabeled membrane-associated coiled coils. We attribute this decrease to competition of the self-association of labeled peptides with unlabeled versions of the same peptides in the core mix. In the FA assay, the opposite effect was seen; signal for these coiled coils increased when assayed with the core mix, but remained the same with the membrane- associated mix. We attribute the increased signal to association between the labeled peptide and peptides in the unlabeled core mix, compared to the signal of the labeled peptide assayed at the low concentration alone or with the membrane-associated mix. We also tested for multi-component association among half-bridge coiled coils. No self- or binary hetero-associations were observed among these coiled coils by CD or pair-wise FRET. FRET was measured for all pairs of fluorophore-labeled HB coiled coils in the presence of a mix of all unlabeled HB coiled coils, and FA was measured for each individual fluorophore-labeled HB coiled coil alone and when mixed with all unlabeled HB coiled coils. In both assays, no increase of signal was observed (FRET data not shown, FA data in Figure 3.6B). Lack of detection may imply that multi-component interactions do not occur, but it is also possible that weak interactions below the level of detection of the assays may occur in the high-density environment of the SPB complex.

Figure 3.5: FRET assay for multi-component interactions. Pairs of fluorophore-labeled orphan coiled coils tested with and without addition of all unlabeled core or membrane- associated coiled coils. Representative data shown are for acceptors and donors with labeled C-termini and were collected after incubating at 25 °C. Error bars represent the standard deviation of 3 independent assays (only one assay for Ndc1 samples). Controls of self-associating Spc29_2 and Spc72_3 included.

Figure 3.6: Fluorescence anisotropy (FA) assay for multi-component interactions. (A) Fluorophore-labeled orphan coiled coils tested with and without addition of all unlabeled core or membrane-associated coiled coils. Representative data shown are for samples labeled with Alexa 568 on the N termini. Error bars represent the standard deviation of 3 independent assays (only one assay for Ndc1 samples). Controls of self-associating Spc29_2 and Spc72_3 included. (B) Fluorophore-labeled half bridge coiled coils tested with and without mix of all unlabeled half bridge coiled coils. Representative data shown are for samples labeled with Alexa 568 on the N termini. No interactions were detected.

Computational annotation predicts few structural features in SPB proteins besides coiled coils. We previously attempted to predict structural features of core SPB proteins using available computational methods (80, 82). Few domains or motifs other than coiled coils were detected by a variety of domain- and fold-recognition programs. The only proteins with predicted structures were Tub4 and Nud1, reported to have a tubulin fold and a leucine-rich repeat fold, respectively. Since then, we extended our analysis to the membrane-associated and half bridge proteins using the same programs and also re-analyzed all SPB proteins with additional programs (3, 4). Again, we found that coiled coils were the predominant motifs predicted within these sequences, although a few additional structural features can be distinguished. The core proteins Spc97 and Spc98 are both predicted with high confidence by Phyre and BioInfoBank (3, 4) to fold into structures in the ARM repeat superfamily. ARM structures are built from repeating units of 3 α- helices that stack into a variety of superhelical, twisted geometries. However, evaluation with HHpred, which has been reported to be a more reliable predictor of this type of alpha-helical repeat fold, did not support this assignment (130, 132). Detection of numerous alpha helices in these long proteins suggests that a tandem repeat structure is likely for Spc97 and Spc98, but the sequence identities are much too low (6-8%), and available fold-recognition methods are not sensitive enough, to assign these proteins to any particular superfamily with confidence. Several proteins from the membrane-associated and half-bridge regions contain transmembrane domains (115-119, 121, 124). Mps2, Mps3, and Kar1 are predicted to have a single transmembrane helix (Figure 3.7). The N-terminal domain of Ndc1 has 5-7 predicted transmembrane helices, depending on what program is used. This domain is predicted to form an α-helical transmembrane bundle, but it does not match any particular superfamily. Particular folds can be predicted only with low confidence for Kar1, the N-terminal domain of Mps2, and the C-terminal domain of Mps3. The predicted secondary structure of Kar1 consists of a repeat pattern of medium-length α-helices alternatively with short-length β-strands. While confidence is low in any particular fold, one common feature among medium-confidence hits from Phyre is a helical extension from an all-alpha or alpha/beta globular domain. The N- terminal domain of Mps2 is predicted with low confidence to form a globular all α-structure, similar to hemoglobin or calmodulin. The C-terminal domain of Mps3 has been characterized by the Winey lab to be a SUN domain that binds a short peptide from Mps2 (133). The structural

70 features of this domain are unknown, but multiple programs predict with low confidence that it forms a globular all-β-structure.

DISCUSSION We have used computational annotations and coiled-coil focused experiments to derive new structural information about the spindle pole body. In this Discussion, we put these new data into perspective by reviewing and integrating known structural information about the spindle pole body into a model that reflects our current understanding. In terms of structural detail, it is clear that much remains to be learned and added to this model. We have resisted the temptation to incorporate speculative inferences not directly supported by experiments. To avoid making unsupported assumptions about how low resolution data melds with higher resolution in the process of model construction, we have integrated information into two different figures. Figure 3.7 illustrates the dimensions of the SPB complex along with sequence features of its constituent proteins. Figure 3.8 contains structural information about the proteins and the interactions between them. Our complementary renderings of the SPB begin to add some details to the prevailing cartoons derived from important early work in this field and can guide future experiments and functional hypotheses.

Figure 3.7: Cross-section of the SPB shows the dimensions and protein composition of each layer or region. The sizes and spacings of the boxes representing the different layers and regions are proportional to the heights and widths of the layers and their spacings, measured in cryo-EM images. Proteins are indicated with colored lines scaled to indicate protein length. Black bars indicate predicted coiled-coil regions and gray bars indicate transmembrane segments. The locations of each protein are approximate and some span multiple layers.

Figure 3.8: Structures, models, and interactions of SPB proteins. SPB proteins are shown using shapes proportional to their primary sequence length or space-filling models corresponding to solved or predicted structures. The proteins are grouped into sub-complexes that are indicated by color: pink – core, blue – membrane-associated, green – half bridge. Coiled coils are represented by long rectangles with a gradient of black to white running from N-terminus to C- terminus. Domains with predicted transmembrane segments are represented as squares. Evidence for interaction between proteins is indicated by lines between shapes, with the type of evidence as shown in the key. Colors/line types are combined where multiple sources of

Overall structure Electron microscopy (EM) studies have provided the overall shape and dimensions of the SPB (Figure 1.1). Its core shape is a cylinder embedded in the nuclear envelope with microtubules attached on either end. The diameter of the complex varies more than the height and is proportional to the number of microtubules attached to the complex. The diameter is ~800-1200 Å in haploid cells and ~1600-1800 Å in diploid cells, while the height is ~1500 Å in both cell types (37-40). The complex consists of multiple layers of differing density. From the most recent and highest resolution studies, five layers can be distinguished and are named as follows: outer plaque (OP), 1st intermediate layer (IL1), 2nd intermediate layer (IL2), central plaque (CP), and inner plaque (IP). The diameter and thickness of each of these layers, along with the distances between the layers, are shown in Figure 3.7 (37). Nuclear microtubules attach to the IP and vary in number according to haplotype to include enough for each chromosome along with a few extra long microtubules that reach toward the opposite SPB (40). Cytoplasmic microtubules attach to the OP and usually number only one tenth of the nuclear microtubules (40). The nuclear envelope aligns with the plane of the CP and has the same approximate thickness (300 Å), but is less dense (37-39). The half bridge extends as an electron-dense rectangular region on one side of the core SPB (38). It appears to be an electron-dense rectangular segment of the nuclear envelope along with a cytoplasmic layer co-planar with IL2 (40). The length of the half bridge co-planar with the nuclear envelope varies with SPB duplication: 600-900 Å for a single SPB and 1200-1500 Å for duplicated SPBs or those in the process of duplication (43, 48). A few cytoplasmic microtubules have been observed attached to the half bridge during G1 of the cell cycle (39). While the SPB does not have an overall symmetry, three pieces of data reflect an apparent symmetry for at least the core cylinder. (1) Electron micrographs revealed that IL2 consists of a hexagonal lattice (37). One protein, Spc42, was sufficient to establish the framework of this lattice, based on its localization to IL2 and the observation that its overexpression led to the formation of the same hexagonal lattice that extends outside the diameter of the SPB (37, 134). (2) The relatively low number of protein types constituting the complex suggests symmetry in effective packing. (3) The SPB approximately doubles in diameter from haploid to diploid cells (38). The difference between these cell types is the number of chromosome and thus the number of microtubules required. It has been proposed that each

microtubule may establish a cross-sectional unit of core proteins through the layers of the cylinder that repeats laterally to generate the appropriate size (46). The number of attached microtubules alone likely does not control the number of units or the size of the SPB. Because a new SPB forms on the half bridge and grows to its full size before insertion into the nuclear envelope and attachment of microtubules (49), control of SPB size is probably at the level of availability of certain proteins. This is supported by the observation that over-expression of CP proteins (Spc110, Spc29, Spc42), but not each individually, can increase the diameter of the CP (127). It is possible that the availability of membrane-associated proteins triggers the end of the growth of a new SPB by binding to the edge of the central plaque and via its transmembrane domains, inserting the new SPB into the nuclear membrane. Work from many labs has identified the proteins that compose the SPB. Below, each of these proteins is described in context of the subcomplex to which it belongs: γ-tubulin complex, core, membrane-associated region, and half bridge. The γ-tubulin complex is technically part of the core cylinder, but it has unique features that are described separately.

γ-tubulin complex The best characterized subcomplex is the γ-tubulin complex, composed of 3 proteins: Tub4, Spc98, Spc97 (45, 47, 112, 128, 135-137). Many copies of this subcomplex populate both ends of the SPB core in the OP and IP. These complexes make the direct connection to the microtubules. The proteins have been shown to associate in a ratio of 2 Tub4: 1 Spc98: 1 Spc97 (112). Recently, an EM structure of the γ-tubulin complex was determined to 25 Å resolution and it reveals a Y shape (Figure 3.8) (47). The proteins have been localized within the Y shape, based on gold labeling of Tub4 and in vivo FRET measurements between each of the proteins. Tub4 resides on the tips of each arm and Spc98 and Spc97 each make up the rest of an arm and part of the stalk. The arms are flexible relative to each other and the authors suggest this flexibility allows Tub4 molecules to come in close proximity, similar to the lateral contacts between Tub1-Tub2 (α-/β-tubulin) heterodimers, but not as close as longitudinal contacts. These dimensions support a template model of microtubule nucleation, with the γ-tubulin complex interacting in a ring at the base of the minus ends of microtubules. A ring complex is seen in higher eukaryotes, but requires more and different types of proteins to form (138, 139). S. cerevisiae does not contain homologs of these associating proteins, but a complex from yeast has been isolated that is the same molecular weight as a ring complex from other organisms and

75 contains multiple γ-tubulin complexes (112). A recent meeting report indicated that the Davis and Agard groups have imaged a ring complex from S. cerevisiae formed in vitro by EM (140), although this is not yet published. Low-resolution information about structure, based on homology and fold recognition, is available for the individual protein components of the γ-tubulin complex. Tub4 is the yeast homolog of human γ-tubulin (39% sequence identity), which forms a helical globular fold, similar to α- and β-tubulin (77). As described in the Results, Spc97 and Spc98 are both likely to form alpha-helical tandem repeats that stack into an elongated shape, which would be consistent with the extended envelope observed for these proteins in the γ-tubulin complex by EM (47).

Core proteins The core of the cylindrical SPB can be resolved into multiple layers by electron microscopy (37, 40). Core proteins have been approximately localized to different layers by immuno-EM or GFP-tagged localization (27, 49). Characterization of which proteins can interact provides further information about the locations of proteins within the complex (36, 46, 49, 127, 129, 141). On the nuclear side of the SPB, the γ-tubulin complex binds to Spc110 and together these proteins make up the IP (112, 135). Spc110 is one of the largest SPB proteins. It is composed of three domains: an N-terminal domain (residues 1-150) that binds the γ-tubulin complex, a central coiled coil (residues 150-800), and a C-terminal domain (residues 800-944) that binds proteins in the central plaque (70, 142). The N-terminal domain binds specifically to Spc98 of the γ-tubulin complex, but no additional details are known about the interaction (112). Two-thirds of the N- terminal domain is predicted to be disordered and may only fold upon interaction with Spc98. The central coiled coil has been shown by EM to form a long, straight rod when expressed alone (70). This rod functions as a spacer between the CP and the IP, as evidenced by the observed decrease in CP-to-IP layer separations in SPBs with coiled coil segments deleted from Spc110 (70). Interestingly, the decrease in inter-layer distance correlates with the amount of coiled-coil sequence removed, but the average distance change per residue is not consistent with a rigid, uninterrupted coiled-coil rod positioned perpendicular to the layers. It may be that the Spc110 coiled coil lies at an angle with respect to this perpendicular, and/or that it exhibits some flexibility, possibly acting as a spring to relieve tension generated by the attached microtubules. These scenarios could also account for the fact that this region has the most variability in layer

thickness between samples in comparison to the other layers. Because coiled-coil deletions of Spc110 mutants were viable, this flexibility or spring function must not be necessary, but perhaps it may help cells grow more efficiently, as these mutant strains grow slowly. Flexibility or spring behavior in Spc110 could be caused by the slight irregularities in the heptad repeat of the coiled coil (143). Coiled coils in myosins have similar irregularities that have been shown to be important for its elastic and flexible properties (144). The CP is consistently the most dense layer of the SPB across different cell fixation and EM preparations (37-40, 126). The proteins packed into this density are the C-terminal domain of Spc110, Cmd1 (the yeast homolog of calmodulin), Spc29, and the N-terminal domain of Spc42. The C-terminal domain of Spc110 and Cmd1 form a tight interaction that has been characterized by a variety of assays (36, 46, 112, 141, 145). Structures of calmodulin from many species are known, including Cmd1 from S. cerevisiae (Figure 3.8) (76). Two EF hands form a binding groove in which a helix can bind. The binding region of Spc110 has been narrowed to residues 900-914 by Y2H (141). These residues are predicted to form an α-helix and contain a 1- 5-8-14 consensus calmodulin-binding motif, where the numbers represent hydrophobic residues and indicate the spacing between them (146). Although the structure of the Cmd1-Spc110 complex has not been solved, these data suggest that the structure is similar to other calmodulins bound to helical peptides. The CP proteins have been shown to interact with each other to varying degrees (36, 46, 49, 112, 127, 129, 141). A model of how these proteins are arranged relative to each other has also been geometrically defined based on in vivo FRET data (46). FRET was measured for pairs of CP proteins (and Cnm67) tagged at either end with a GFP variant (FP) and the FRET signal strength was classified into bins of relative distances. These distance bins were converted to constraints between the different protein pairs and the authors found that a single, unique geometry of the proteins was consistent with their data. The FRET-based geometry was repeated with hexagonal symmetry to construct a model that satisfied all of the constraints. This is the most detailed model available so far for any extensive portion of the SPB. In addition to relative locations of proteins, the model includes information about orientation that was obtained from differential measurement of FRET with FP on either end of nearly all proteins tested. We suggest that one minor modification to the model could be made to better explain patterns in the viability of different yeast strains observed in the Muller et al. study (46).

Detecting FRET signal depended on the viability of yeast strains, each harboring FP-tagged protein alone or a pair of tagged proteins. All strains carrying single FP-tagged proteins were viable, but combinations of FP-Spc29 with Spc110-FP and FP-Spc29 with FP-Spc42 were not (abbreviations indicate tagging of the N-terminus (FP-x) or C-terminus (x-FP)). This may be significant, if the FP fusions are preventing interactions that are necessary for the integrity and function of the SPB. This would suggest that the labeled termini in the inviable strains lie in close proximity in a functional SPB. The model of the core SPB was devised based only on measured FRET signal, so these inviable pairs and any information they can convey were not considered in the model. Observed FRET signal between Spc29-FP with Spc110-FP and Spc29- FP with FP-Spc42 indicated placement in the model of the C-terminus of Spc29 near the C- terminus of Spc110 and the N-terminus of Spc42. Simply flipping the orientation of Spc29 would make the data more consistent with the viability observations, although the inviable pairs may be in closer proximity than shown in the model. The majority of Spc42 is predicted to form 3 different coiled coils, with disordered regions predicted for the first and last 50 residues and in between the coiled-coil regions. Only the first coiled-coil region has been confirmed to be a parallel homodimeric coiled coil via X-ray crystallography and biochemical solution studies (129). Two coiled coils from Spc29 are predicted to encompass half of Spc29 sequence. The second coiled-coil region has been confirmed to form a homotrimeric coiled coil via biochemical solution studies (129). The other putative coiled coils from Spc42 and Spc29 do not homo-associate in isolation and have not been observed to interact with each other or with any other putative coiled coil in the core or membrane-associated proteins. These regions may still form some α-helical structure, even a coiled coil, but any such structures are not stable under the conditions tested, without the rest of the proteins or SPB complex. These “orphan” coiled-coil regions may also homo-oligomerize if stabilized by the strongly self-associating coiled-coil regions elsewhere in the sequence mentioned above. They may also form weak hetero-associations with each other or with other coiled coils, below the sensitivity of our assays. Such interactions could be important in the assembly or polymerization of units into the larger SPB complex. The resolution of experiments localizing proteins to particular layers by immuno-EM and/or GFP fusions does not allow exact placement of proteins within the complex. In particular, the intermediate layers (IL1 and IL2) were not observed in early EM studies of the SPB (38, 39).

Cnm67 and Spc42 are placed in these layers in our model with defined orientation based on the following reported interactions. The N-terminal domain of Spc42 interacts only with CP proteins, whereas the C-terminal domain of Spc42 interacts with the N-terminal domain of Cnm67 in the IL1/2 region. The C-terminal domain of Cnm67 interacts only with OP (46, 49, 127). No other proteins are known to reside in IL1 or IL2 and no other interactions have been reported between proteins of the CP and OP, although thorough investigation of all possible protein pairs has not been executed. Additional evidence placing Spc42 in IL2 comes from overexpression studies of Spc42, which resulted in an extension of IL2 on top of the cytoplasmic side of the nuclear envelope (37). The coiled-coil regions of Spc42 and Cnm67 may play roles as spacers, similar to the way the coiled-coil region of Spc110 separates the CP and IP (70, 71). The first coiled-coil region of Spc42 connects CP to IL2, although the coiled coil may extend into one or both of these layers since the spacing between them is less than the length of this coiled- coil segment (100 Å) (129). It is clear that the coiled-coil region of Cnm67 forms a spacer between IL2 and OP, because the IL2-to-OP separation has been shown to decrease or increase when this region of the protein has been shortened or lengthened. This observation leads to the conclusion that the layer in between, IL1, consists of the coiled-coil region of Cnm67. This layer is denser than the regions where the coiled coils of Spc42 and Spc110 reside (37, 40). This extra density may be caused by tighter packing of the Cnm67 coiled coil or by associating proteins. Components of the OP are the N-terminal domain of Cnm67, Nud1, Spc72, and the γ- tubulin complex. Nud1 interacts with both Cnm67 and Spc72 via its C-terminal domain (36, 147). This domain is predicted with high confidence to fold into a leucine-rich repeat (LRR), similar to toll-like receptors (TLR) (Figure 3.8). Spc72 has been shown to interact via its N- terminal domain with both Spc97 and Spc98 and via its C-terminal domain with Nud1 (136, 147, 148). Spc72 has three coiled-coil regions, each shown to self-associate to varying degrees (129). The third coiled coil self-associates the strongest and was characterized via solution assays to be a parallel homodimer. The second coiled coil is rather long and self-associates weakly. The first coiled coil shows evidence of self-association and hetero-association with the second coiled coil as well as the coiled coil from Spc97. Interestingly, despite its many and long coiled-coil domains, Spc72 has not been observed to serve as a spacer between layers. Rather, it resides in an electron-dense layer. One possible model is that the protein forms a more compact structure,

via incorporation of coiled-coil segments interspersed with non-coiled-coil regions into a denser, globular structure.

Membrane-associated proteins Four proteins are associated with the nuclear envelope near the SPB: Ndc1, Mps2, Bbp1, and Nbp1(43, 44, 117, 119). Despite association with the nuclear membrane, only Ndc1 and Mps2 have predicted transmembrane regions. The other proteins are likely localized through their interactions with these two proteins. Ndc1 interacts with Nbp1 and Mps2 interacts with both Bbp1 and Nbp1 (44, 117, 149). The N-terminal domain of Ndc1 has multiple transmembrane helices that likely form an α-helical bundle similar to membrane transporters or ion channels. No other domains or structures are predicted for the C-terminal domain of Ndc1, except a coiled coil in the middle of the protein for which we have not been able to find an interacting partner. This coiled coil may be a false positive prediction or it may only form in the context of the rest of the protein or complex. Prior studies have localized the interaction between Mps2 and Bbp1 to the N-terminal two-thirds of Mps2 and the C-terminal half of Bbp1 (117). Now we have classified the interaction as a hetero-associating coiled coil, likely with an ~ 1:4 ratio of Bbp1 to Mps2. Each protein also has a second coiled-coil domain that has been shown to homo-dimerize. These coiled coils account for approximately half of each of these proteins. While structural features for the rest of these proteins are unknown, both Mps2 and Bbp1 proteins are scaffold or crosslinking proteins because they have been observed to interact with two other proteins each and are likely tethered to the nuclear membrane via Mps2’s transmembrane domain. Some of these membrane-associated proteins must interact with one or more proteins within the CP of the core to allow insertion of the core cylinder into the nuclear membrane, although the interactions may be relatively weak. The only documented interaction is between Bbp1 and Spc29, but details are unknown (117). We tested whether the interaction was occurring via non-self-associating putative coiled-coil regions, but we did not observe any evidence of interaction. The interaction between Bbp1 and Spc29 is likely to be through non-coiled-coil regions, especially given that both of Bbp1’s coiled coils have other known partners.

Half-bridge proteins Four proteins have been localized to the half bridge: Kar1, Sfi1, Cdc31, and Mps3. All of the proteins have been localized to the cytoplasmic side of the half bridge, but Mps3 also resides on the nuclear side (121-124). This localization difference is confirmed by the greater electron density of the half bridge on the cytoplasmic side, and it may direct SPB duplication, since a new SPB assembles specifically on the cytoplasmic side (38, 49). The only interactions that have been demonstrated between half-bridge proteins are interactions of Cdc31 with each of the other three components (though interaction with Mps3 is weaker than others and so is not featured in Figure 3.8) (44, 48, 78, 120-122, 124, 150). However, the only other possible interaction that has been tested, or at least reported, is between Mps3 and Kar1; these proteins did not interact in a Far Western assay (121). Cdc31 is the yeast homolog of centrin, which folds into a similar structure as Cmd1/calmodulin (78). Like calmodulin, the two EF hands of Cdc31 form a binding groove to which a peptide can bind. The Cdc31 binding regions of Kar1 and Sfi1 have been identified, and structures of both have been determined, by NMR and X-ray crystallography, respectively (48, 78). A helical peptide from the middle of Kar1 binds to a groove in Cdc31 (Figure 3.8). Sfi1 has multiple consensus repeat domains, five of which are featured in two crystal structures where each repeat is bound to an independent copy of Cdc31 (Figure 3.8). A model of Sfi1 orients a globular, mostly helical N-terminal domain toward the membrane-associated proteins, a long α- helix of repeat domains, binding up to 21 copies of Cdc31, parallel to the plane of the membrane, and a disordered C-terminal domain towards the middle of a full bridge (48). Kar1 and Mps3 have multiple putative coiled-coil regions. Much to our surprise, none of the coiled coils from Kar1 or Mps3 were observed to self-associate or hetero-associate with each other. The coiled coils may need the rest of the proteins for stability, which could arise via concentration at the nuclear membrane by their transmembrane domains. The coiled coils in Kar1 may be false-positive prediction errors, as Kar1 is predicted to form a globular fold of long α-helices and short β-strands. Apart from the coiled-coil and transmembrane domains, Mps3 also contains a SUN domain that is predicted to form a globular domain of β-strands (44). Only two interactions between half-bridge and membrane-associated proteins have been observed, despite many combinations tested. Kar1 and Bbp1 have been shown to bind via Y2H but not coIP (117). Mps2 and Mps3 have been shown to bind via the SUN domain of Mps3 and

the last 15 residues of Mps2 (44). Interactions between core and HB proteins would presumably occur between a newly duplicated SPB and the cytoplasmic side of the HB. The proteins detected at a newly duplicated SPB are Spc42, Spc29, Nud1, and Cnm67 (49), but interactions with these proteins and HB proteins have not been observed. The interactions may be too weak for detection or may be mediated by another unknown protein. A few interactions have been observed between half-bridge and core proteins. Cdc31 has been shown to bind the C-terminal domain of Spc110 in vitro, but this interaction may be the same binding site as Cmd1 and thus possibly not relevant in vivo (150). A portion of Spc72 has been localized to the half bridge during G1 of the cell cycle (49). Its C-terminal domain has been observed to bind to the middle of Kar1 (127). This interaction allows attachment of cytoplasmic microtubules to the half bridge during G1. Spc72 and the cytoplasmic microtubules have been observed through the entire half bridge, whereas a newly duplicated SPB has only been observed on the distal tip of the half bridge (38, 49), so the interaction of Spc72 with Kar1 is functionally distinct from the newly duplicated SPB.

Summary and future directions The data described above are pieces of the puzzle that can be assembled into a model of the SPB structure. While some pieces are missing, making a model from what is known can help direct new studies to complete the picture. We have incorporated the dimensions of the complex and the various layers or sub-complexes, the identities and approximate locations of the proteins, known or predicted structural features of the proteins, and interactions between the proteins. We also include our own work characterizing predicted coiled coils and their interactions to generate models of the overall structure that, when taken together, are as complete as possible (Figures 3.7 and 3.8). The models immediately reveal areas of future investigations: (1) Structural analysis of the SPB coiled coils could be continued by solving crystal structures of the identified coiled coils and finding partners for the orphan coiled coils with other assays and constructs. (2) Only a rough idea of the stoichiometry of SPB proteins is known, but quantification of each protein is critical for populating a model of the complex. All core proteins are found in approximately equal quantities, with the exception of Tub4 and Spc42, which are present at approximately twice this level, based on SDS-PAGE of SPB purifications (27). Spc110 may also be present in 2-fold excess, but there are conflicting reports (27, 46). Even less is known about the numbers of

membrane-associated and half-bridge proteins, but based on the crystal structure of Cdc31 and Sfi1, the ratio between them is likely ~21 Cdc31: 1 Sfi1(48). (3) Testing all possible interactions between the proteins, and ideally identifying the interacting regions, would give a more thorough protein interaction network and help place proteins within the complex with more certainty. (4) While determining the structure of each protein would be a beneficial but formidable task, the shape of each protein may be obtained more easily through gel filtration or density gradients. This information would help to determine how the proteins may fit together within the complex. (5) A higher resolution model could be obtained with single particle EM, using technical advances in the field. Symmetry of the complex could be probed through correlating quantitative information between the size and shape of the SPB and sub-complexes and layers along with the number of microtubules attached. (6) This additional data may be integrated with current structural information using computational methods, similar to recent efforts with the nuclear pore complex (13), to produce a 3D self-consistent model of the SPB with computational methods. Thus, despite the many technical challenges that confront structural studies of the SPB, an integrated approach of deconstruction and re-assembly, illustrated here focusing on the SPB coiled-coil motifs, promises new insights.

CHAPTER 4

Conclusions and Future Directions

Conclusions In my thesis, I have described my efforts to characterize coiled coils in the spindle pole body (SPB) complex. Coiled coils had been recognized within SPB protein sequences for a long time, but most had not been experimentally validated. I found that many SPB coiled coils self- associate and several coiled coils form hetero-associations. These data contribute constraints as to how SPB proteins interact within the complex. I constructed models of the SPB structure by incorporating the coiled-coil data with other structural information about the SPB. Many large protein complexes, like the SPB, are intractable for structural determination. Incorporation of multiple types of indirect structural information, such as the extensive characterization of coiled-coil interactions that I performed, can help build a model of a complex. While a three-dimensional (3D) model of the SPB cannot be generated currently, sufficient data exist to model some of the complex at different resolutions. My models in Chapter 3 are at relatively low resolution, but describe all of the proteins. Other models describe subsets of SPB proteins at higher resolution. The interaction between the microtubules and the γ-tubulin complex has been modeled based on separate EM images of the microtubules and the γ-tubulin complex and crystal structures of their components (47). The models from Muller et al. describing placement of the core proteins may be expanded to a 3D model with some assumptions of the proteins’ sizes and shapes (46). Determining how proteins interact at atomic-level detail is important for understanding how the proteins assemble into complexes. Identifying which proteins interact is relatively easy, especially with assays adapted for high-throughput performance, but determining how the interactions occur is more difficult. Many efforts have been underway to predict computationally how proteins interact, but the programs rely on inference from known structures of proteins or domains in complex, and are rather error prone (151, 152). Good methods exist to predict the propensity of a sequence to form a coiled coil (61-64), but it is much more difficult to predict the precise details of the complex structure. In particular, predicting the correct interaction partner

for a putative coiled-coil helix is an unsolved challenge (despite progress predicting coiled-coil interactions for bZIP transcription factors) (153). Experimental characterization is still necessary to reveal structural features of coiled-coil interactions. From the analysis of SPB coiled coils presented in this thesis, I am struck by the large number of orphan SPB coiled coils and the small number of hetero-associations between SPB coiled coils revealed by my experiments. While several methods were used to detect interaction among putative SPB coiled coils, I was unable to identify partners for many of them. These peptide sequences may not form coiled coils, despite resemblance to other coiled coils that allowed detection by algorithms. Alternatively, the appropriate coiled-coil interactions for the orphans may have been too weak for detection by the assays or they may require other parts of the protein to form. For example, the orphan coiled coils in Spc42 and Spc29 may homo- oligomerize, but only when stabilized by the strongly self-associating coiled-coil region located elsewhere in each protein’s sequence. The observation that all of the putative coiled coils in both Kar1 and Mps3 are orphans suggests that they may form intra-molecular coiled-coil interactions that contribute to the overall structure of these proteins, but require stabilization from non-coiled- coil regions. The high-density environment of the SPB could also contribute to coiled-coil formation. Helices from coiled coils may homo-oligomerize or hetero-oligomerize and these oligomerization preferences cannot be discerned by sequence as yet. The large size of the SPB requires extensive oligomerization, involving homo- and hetero-association of multiple copies of each protein. Coiled coils were expected to play a role in this oligomerization by creating homo- oligomers, by mediating interaction between different proteins, or by forming intra-molecularly as part of a protein’s structure. My data show that the first role dominates the observed SPB coiled-coil interactions. Homo-oligomeric coiled coils from Spc110, Cnm67, and Spc42 likely form support linkers between electron-dense layers, but others may help form structural units that repeat throughout the complex. For a full picture of the SPB, more structural information must be obtained. Some possible future directions are described at the end of Chapter 3. Here I propose additional experiments that focus on expanding structural analysis of the SPB coiled coils from my current findings. Other large complexes with coiled-coil proteins, such as the centrosome and

kinetochore, could also benefit from similar characterization of their coiled coils to provide structural information.

Future Directions

Additional characterization of SPB coiled coils I identified a number of coiled-coil interactions among SPB proteins, and I determined some of their structural features, including stoichiometry, partnering, and orientation. Structural features yet to be examined include the extent of coiled-coil formation and the alignment of the helices within the coiled coils. Although the presence of a coiled coil in a protein sequence can be predicted well with computational algorithms, the boundaries of the coiled coils cannot. I manually inspected the sequence of each SPB coiled coil and chose to produce peptides that corresponded to coiled-coil regions with generous cutoffs. With this choice, I expected that potentially only part of each peptide may form a stable coiled-coil interaction. In fact, all of the self-associating coiled coils showed evidence of only partly forming coiled coils, based on helical content determined by CD (Tables 2.2 and 3.1). The extent of stable coiled-coil formation could further be determined by crosslinking, proteolysis, and mass spectrometry experiments. These assays may also be used to determine the axial alignment between helices of the coiled coils. The most stable alignment often minimizes offset of helices to maximize shielding of core hydrophobic residues, but may also depend on specific residue interactions. Most parallel homo- oligomers have Cn symmetry with the same residues from each strand interacting in the core. Coiled coils with other properties, especially hetero-oligomers of helices with different lengths, may have more offset helices. For a coiled coil with successful characterization of all of these structural features, a reliable structural model may be constructed. Alternatively, structural features can be fully determined by solving the crystal structure of each coiled coil. Crystallization of each self-associating SPB coiled coil was attempted, but proved unsuccessful for most of them. One coiled coil from Spc72 (Spc72_3) did crystallize and diffracted to 2.4 Å. Because this coiled coil was determined to be a parallel dimer in solution, I tried to solve the structure with molecular replacement using a set of parallel coiled-coil dimers as starting models. However, I was unable to find a model that fit the data well. One possible solution for moving forward with all of these self-associating SPB coiled coils is modifying the peptide constructs. To help with Spc72_3 phasing, a serine predicted to be on the surface of the 87

coiled coil could be replaced with a cysteine. This cysteine would provide a mercury-binding site for multiple isomorphous replacement. Coiled coils that did not crystallize could be shortened to the minimal folded/interacting region, based on proteolysis analysis or very high scores from coiled-coil prediction algorithms. Alternatively, they could be lengthened by attaching a canonical coiled coil on one end. Crystallization of many parallel coiled coils has been aided by an in-frame fusion with dimeric GCN4 or its trimeric variants (154, 155). In the discussions of Chapters 2 and 3, models were proposed for particular proteins that contain self-associating coiled coils. The coiled-coil region of Cnm67 has been suggested to form a coiled-coil support separating electron-dense layers (71). I tested the region in shorter constructs due to breaks in the coiled-coil prediction, but my data suggest that the entire coiled- coil region forms a parallel homo-oligomer. Spc72 contains three coiled coils that are significantly separated in sequence, but were all shown to self-associate. The first and last coiled coils were shown to form dimers, thus it is likely that the entire protein dimerizes. Unlike the coiled-coil region of Cnm67 and other SPB coiled coils that form supports between layers, Spc72 resides in an electron-dense layer, and may form an interesting globular domain with both coiled-coil and non-coiled coil regions. These models of Cnm67 and Spc72 could be tested by structural determination of individual proteins or constructs containing all of the coiled-coil regions by X-ray crystallography or cryo-EM. Structure determination or extensive characterization of particular pairs or subsets of proteins together would be useful in constructing the SPB model. Good targets for this approach are the hetero-associating coiled coils, either isolated or in the context of the rest of the protein. Characterization of the hetero-association between coiled coils in Bbp1 and Mps2 suggests that the coiled coils form a hetero-pentamer, although this finding requires further confirmation. Structural examples of hetero-pentameric coiled coils currently do not exist, so a structure of this complex would be informative for both the SPB and coiled-coil fields. Crystallization trials of this complex could be set up soon with the tagless, truncated constructs that were used to confirm the hetero-association by CD. In addition, characterization of the whole proteins in complex is appealing, although may be more difficult. Each protein also has a second coiled-coil domain that has been shown to homo-dimerize. It is not clear whether the coiled coils would still form in the stoichiometry that was characterized in isolation and if so, how the coiled coils are accommodated within the structure of the interacting proteins.

Two of the SPB hetero-associations share a common partner; a coiled coil from Spc72 (Spc72_1) can hetero-associate with a different coiled coil from Spc72 (Spc72_2) or with the coiled coil from Spc97. One way to determine which coiled-coil interaction occurs within the SPB is to characterize how Spc72 interacts with the γ-tubulin complex that includes Spc97. The γ-tubulin complex has been structurally characterized with single particle cryo-EM (47). The same method and conditions could be used with a purified mixture of the γ-tubulin complex and Spc72. The resulting cryo-EM images could be compared to images of the γ-tubulin complex alone to determine a low-resolution structure of Spc72 and how it interacts with the γ-tubulin complex. Many putative SPB coiled coils remain orphans; I was unable to find a partner peptide to allow coiled-coil formation. It is not clear whether they are false-positive prediction errors, whether the partner helix resides in a protein outside of the SPB complex, or if these sequences require the rest of the protein to stabilize the coiled-coil interactions. To address the latter possibility, each SPB protein with orphan coiled coils could be expressed and tested as the whole protein. They could be tested for self-association or interaction with the isolated SPB coiled-coil peptides already generated. Many SPB proteins have been observed to be phosphorylated in vivo (27, 71, 136, 147, 156, 157). The role of phosphorylation in SPB assembly and function is not well understood, but particular proteins (Spc42, Spc29, Spc110) have been shown to require phosphorylation for proper assembly into the SPB (156-158). The kinases Cdc28 and Mps1 have been shown to be responsible for SPB phosphorylation and specific phosphorylated sites have been located in a few SPB proteins (Spc42, Spc29) (156, 158, 159). While these sites do not occur within predicted coiled-coil regions of these proteins, it is possible that other residues may be phosphorylated within some coiled coils and that the phosphorylation regulates coiled-coil formation. This hypothesis could be tested in vitro by examining whether isolated putative coiled-coil peptides are phosphorylated by Cdc28 and/or Mps1 and then assessing homo- or hetero-oligomerization of phosphorylated peptides with the assays I have already developed. The significance of coiled coils in SPB proteins in vivo could be investigated through disruption, stabilization, or replacement of the coiled-coil interactions that I have identified. Mutations to core residues of a coiled coil to charged, proline, or glycine residues has been shown to prevent coiled-coil formation (60), and these mutations in SPB coiled coils may result

in disruption of SPB complex formation. Many temperature-sensitive mutations in Spc42 have been mapped to the first coiled-coil domain and these mutants have abnormal, defective SPBs (134). Stabilizing mutations in SPB coiled coils may also disrupt SPB assembly or function. For example, in tropomyosin, coiled-coil stabilization has been achieved by mutating unusual residues in the core to canonical coiled-coil residues, which resulted in disruption of tropomyosin’s abililty to bind to actin (160). Replacing the SPB coiled coils with coiled coils with similar structural features could test whether the coiled coils are simple, modular interaction units or whether they have sequence-specific features that contribute to SPB assembly.

Coiled coils in other complexes Coiled coils have been predicted in proteins from many other complexes. Experimental characterization of most of these coiled coils has not been achieved, but could be helpful for giving structural information about the complexes. I recommend additional studies for two complexes in particular, the centrosome and the kinetochore, based on their relation to the SPB. The centrosome has an overall shape very different from the SPB, and is composed mostly of proteins non-homologous to SPB proteins (35, 50, 51). Despite these differences, most of the proteins from both complexes feature one or more predicted coiled coils. The coiled coils in centrosomal proteins may perform similar roles as in SPB proteins. The majority of SPB coiled coils were found to homo-oligomerize as parallel dimers, some forming spacers between electron-dense layers. Centrosomal coiled coils may also form homo-oligomeric modules. However, differences in the types of coiled-coil interactions between the two complexes may account for and affect the shape of the two complexes. For example, the centrosome does not have distinct electron-dense layers, so centrosomal coiled coils may not function as spacer supports. This hypothesis could be tested by experimentally characterizing isolated centrosomal coiled coils. Kinetochore proteins are also predicted to contain many coiled coils (24). Interestingly, the spindle pole body shares greater similarity in overall shape to the yeast kinetochore than to the centrosome. The kinetochore is a cylinder, but it binds to a single microtubule and consists of three layers: two electron-dense layers with a spacer region in between. In one characterized kinetochore sub-complex, coiled coils from four proteins appear to span this spacer region, similar to SPB coiled coils that span spacer regions (24). However, unlike the SPB coiled coils forming homo-dimers, the kinetochore coiled coils form two hetero-dimers that partially overlap

90 as a hetero-tetramer through the spacer region. Some other proteins with putative coiled coils have not been localized to particular regions of the kinetochore. Experimental characterization of kinetochore coiled coils could help place proteins within the complex and reveal structural aspects of the kinetochore.

I have demonstrated how characterizing coiled-coil interactions between proteins within a complex can contribute structural information about the complex. Continued experimental characterization of native coiled coils will not only increase understanding of the structural aspects and function of this prevalent motif, but also provide additional insight into the organization of large protein complexes, such as the SPB.

REFERENCES (1) Alber, F., Forster, F., Korkin, D., Topf, M., and Sali, A. (2008) Integrating diverse data for structure determination of macromolecular assemblies. Annu Rev Biochem 77, 443- 77. (2) Bernstein, F. C., Koetzle, T. F., Williams, G. J., Meyer, E. F., Jr., Brice, M. D., Rodgers, J. R., Kennard, O., Shimanouchi, T., and Tasumi, M. (1977) The Protein Data Bank: a computer-based archival file for macromolecular structures. J Mol Biol 112, 535-42. (3) Kajan, L., and Rychlewski, L. (2007) Evaluation of 3D-Jury on CASP7 models. BMC Bioinformatics 8, 304. (4) Kelley, L. A., and Sternberg, M. J. (2009) Protein structure prediction on the Web: a case study using the Phyre server. Nat Protoc 4, 363-71. (5) Punta, M., and Ofran, Y. (2008) The rough guide to in silico function prediction, or how to use sequence and structure information to predict protein function. PLoS Comput Biol 4, e1000160. (6) Schwede, T., Sali, A., Honig, B., Levitt, M., Berman, H. M., Jones, D., Brenner, S. E., Burley, S. K., Das, R., Dokholyan, N. V., Dunbrack, R. L., Jr., Fidelis, K., Fiser, A., Godzik, A., Huang, Y. J., Humblet, C., Jacobson, M. P., Joachimiak, A., Krystek, S. R., Jr., Kortemme, T., Kryshtafovych, A., Montelione, G. T., Moult, J., Murray, D., Sanchez, R., Sosnick, T. R., Standley, D. M., Stouch, T., Vajda, S., Vasquez, M., Westbrook, J. D., and Wilson, I. A. (2009) Outcome of a workshop on applications of protein models in biomedical research. Structure 17, 151-9. (7) Zhang, Y. (2009) Protein structure prediction: when is it useful? Curr Opin Struct Biol 19, 145-55. (8) Stahlberg, H., and Walz, T. (2008) Molecular electron microscopy: state of the art and current challenges. ACS Chem Biol 3, 268-81. (9) Droit, A., Poirier, G. G., and Hunter, J. M. (2005) Experimental and bioinformatic approaches for interrogating protein-protein interactions to determine protein function. J Mol Endocrinol 34, 263-80. (10) Piehler, J. (2005) New methodologies for measuring protein interactions in vivo and in vitro. Curr Opin Struct Biol 15, 4-14. (11) Gorba, C., Miyashita, O., and Tama, F. (2008) Normal-mode flexible fitting of high- resolution structure of biological molecules toward one-dimensional low-resolution data. Biophys J 94, 1589-99. (12) Trabuco, L. G., Villa, E., Mitra, K., Frank, J., and Schulten, K. (2008) Flexible fitting of atomic structures into electron microscopy maps using molecular dynamics. Structure 16, 673-83. (13) Alber, F., Dokudovskaya, S., Veenhoff, L. M., Zhang, W., Kipper, J., Devos, D., Suprapto, A., Karni-Schmidt, O., Williams, R., Chait, B. T., Rout, M. P., and Sali, A. (2007) Determining the architectures of macromolecular assemblies. Nature 450, 683-94. (14) Alber, F., Dokudovskaya, S., Veenhoff, L. M., Zhang, W., Kipper, J., Devos, D., Suprapto, A., Karni-Schmidt, O., Williams, R., Chait, B. T., Sali, A., and Rout, M. P. (2007) The molecular architecture of the nuclear pore complex. Nature 450, 695-701. (15) Cheng, Y. (2009) Toward an atomic model of the 26S proteasome. Curr Opin Struct Biol 19, 203-8. (16) Murata, S., Yashiroda, H., and Tanaka, K. (2009) Molecular mechanisms of proteasome assembly. Nat Rev Mol Cell Biol 10, 104-15.

(17) Groll, M., Ditzel, L., Lowe, J., Stock, D., Bochtler, M., Bartunik, H. D., and Huber, R. (1997) Structure of 20S proteasome from yeast at 2.4 A resolution. Nature 386, 463-71. (18) Unno, M., Mizushima, T., Morimoto, Y., Tomisugi, Y., Tanaka, K., Yasuoka, N., and Tsukihara, T. (2002) The structure of the mammalian 20S proteasome at 2.75 A resolution. Structure 10, 609-18. (19) da Fonseca, P. C., and Morris, E. P. (2008) Structure of the human 26S proteasome: subunit radial displacements open the gate into the proteolytic core. J Biol Chem 283, 23305-14. (20) Nickell, S., Beck, F., Scheres, S. H., Korinek, A., Forster, F., Lasker, K., Mihalache, O., Sun, N., Nagy, I., Sali, A., Plitzko, J. M., Carazo, J. M., Mann, M., and Baumeister, W. (2009) Insights into the molecular architecture of the 26S proteasome. Proc Natl Acad Sci U S A 106, 11943-7. (21) Nickell, S., Mihalache, O., Beck, F., Hegerl, R., Korinek, A., and Baumeister, W. (2007) Structural analysis of the 26S proteasome by cryoelectron tomography. Biochem Biophys Res Commun 353, 115-20. (22) Cheeseman, I. M., and Desai, A. (2008) Molecular architecture of the kinetochore- microtubule interface. Nat Rev Mol Cell Biol 9, 33-46. (23) Santaguida, S., and Musacchio, A. (2009) The life and miracles of kinetochores. Embo J 28, 2511-31. (24) Welburn, J. P., and Cheeseman, I. M. (2008) Toward a molecular structure of the eukaryotic kinetochore. Dev Cell 15, 645-55. (25) Cheeseman, I. M., Niessen, S., Anderson, S., Hyndman, F., Yates, J. R., 3rd, Oegema, K., and Desai, A. (2004) A conserved protein network controls assembly of the outer kinetochore and its ability to sustain tension. Genes Dev 18, 2255-68. (26) Meraldi, P., McAinsh, A. D., Rheinbay, E., and Sorger, P. K. (2006) Phylogenetic and structural analysis of centromeric DNA and kinetochore proteins. Genome Biol 7, R23. (27) Wigge, P. A., Jensen, O. N., Holmes, S., Soues, S., Mann, M., and Kilmartin, J. V. (1998) Analysis of the Saccharomyces spindle pole by matrix-assisted laser desorption/ionization (MALDI) mass spectrometry. J Cell Biol 141, 967-77. (28) Joglekar, A. P., Bloom, K., and Salmon, E. D. (2009) In vivo protein architecture of the eukaryotic kinetochore with nanometer scale accuracy. Curr Biol 19, 694-9. (29) Brohawn, S. G., Partridge, J. R., Whittle, J. R., and Schwartz, T. U. (2009) The nuclear pore complex has entered the atomic age. Structure 17, 1156-68. (30) Elad, N., Maimon, T., Frenkiel-Krispin, D., Lim, R. Y., and Medalia, O. (2009) Structural analysis of the nuclear pore complex by integrated approaches. Curr Opin Struct Biol 19, 226-32. (31) Devos, D., Dokudovskaya, S., Alber, F., Williams, R., Chait, B. T., Sali, A., and Rout, M. P. (2004) Components of coated vesicles and nuclear pore complexes share a common molecular architecture. PLoS Biol 2, e380. (32) Schwartz, T. U. (2005) Modularity within the architecture of the nuclear pore complex. Curr Opin Struct Biol 15, 221-6. (33) Beck, M., Forster, F., Ecke, M., Plitzko, J. M., Melchior, F., Gerisch, G., Baumeister, W., and Medalia, O. (2004) Nuclear pore complex structure and dynamics revealed by cryoelectron tomography. Science 306, 1387-90. (34) Kampmann, M., and Blobel, G. (2009) Three-dimensional structure and flexibility of a membrane-coating module of the nuclear pore complex. Nat Struct Mol Biol 16, 782-8.

(35) Jaspersen, S. L., and Winey, M. (2004) The budding yeast spindle pole body: structure, duplication, and function. Annu Rev Cell Dev Biol 20, 1-28. (36) Knop, M., Pereira, G., and Schiebel, E. (1999) Microtubule organization by the budding yeast spindle pole body. Biol Cell 91, 291-304. (37) Bullitt, E., Rout, M. P., Kilmartin, J. V., and Akey, C. W. (1997) The yeast spindle pole body is assembled around a central crystal of Spc42p. Cell 89, 1077-86. (38) Byers, B., and Goetsch, L. (1974) Duplication of spindle plaques and integration of the yeast cell cycle. Cold Spring Harb Symp Quant Biol 38, 123-31. (39) Byers, B., and Goetsch, L. (1975) Behavior of spindles and spindle plaques in the cell cycle and conjugation of Saccharomyces cerevisiae. J Bacteriol 124, 511-23. (40) O'Toole, E. T., Winey, M., and McIntosh, J. R. (1999) High-voltage electron tomography of spindle pole bodies and early mitotic spindles in the yeast Saccharomyces cerevisiae. Mol Biol Cell 10, 2017-31. (41) Helfant, A. H. (2002) Composition of the spindle pole body of Saccharomyces cerevisiae and the proteins involved in its duplication. Curr Genet 40, 291-310. (42) Rout, M. P., and Kilmartin, J. V. (1990) Components of the yeast spindle and spindle pole body. J Cell Biol 111, 1913-27. (43) Winey, M., Baum, P., Goetsch, L., and Byers, B. (1991) Genetic determinants of spindle pole body duplication in budding yeast. Cold Spring Harb Symp Quant Biol 56, 705-8. (44) Araki, Y., Lau, C. K., Maekawa, H., Jaspersen, S. L., Giddings, T. H., Jr., Schiebel, E., and Winey, M. (2006) The Saccharomyces cerevisiae spindle pole body (SPB) component Nbp1p is required for SPB membrane insertion and interacts with the integral membrane proteins Ndc1p and Mps2p. Mol Biol Cell 17, 1959-70. (45) Sobel, S. G., and Snyder, M. (1995) A highly divergent gamma-tubulin gene is essential for cell growth and proper microtubule organization in Saccharomyces cerevisiae. J Cell Biol 131, 1775-88. (46) Muller, E. G., Snydsman, B. E., Novik, I., Hailey, D. W., Gestaut, D. R., Niemann, C. A., O'Toole E, T., Giddings, T. H., Jr., Sundin, B. A., and Davis, T. N. (2005) The organization of the core proteins of the yeast spindle pole body. Mol Biol Cell 16, 3341- 52. (47) Kollman, J. M., Zelter, A., Muller, E. G., Fox, B., Rice, L. M., Davis, T. N., and Agard, D. A. (2008) The structure of the gamma-tubulin small complex: implications of its architecture and flexibility for microtubule nucleation. Mol Biol Cell 19, 207-15. (48) Li, S., Sandercock, A. M., Conduit, P., Robinson, C. V., Williams, R. L., and Kilmartin, J. V. (2006) Structural role of Sfi1p-centrin filaments in budding yeast spindle pole body duplication. J Cell Biol 173, 867-77. (49) Adams, I. R., and Kilmartin, J. V. (1999) Localization of core spindle pole body (SPB) components during SPB duplication in Saccharomyces cerevisiae. J Cell Biol 145, 809- 23. (50) Luders, J., and Stearns, T. (2007) Microtubule-organizing centres: a re-evaluation. Nat Rev Mol Cell Biol 8, 161-7. (51) Azimzadeh, J., and Bornens, M. (2007) Structure and duplication of the centrosome. J Cell Sci 120, 2139-42. (52) Bornens, M. (2002) Centrosome composition and microtubule anchoring mechanisms. Curr Opin Cell Biol 14, 25-34. (53) Nigg, E. A. (2004) Centrosomes in development and disease, Wiley-VCH, Weinheim.

(54) Andersen, J. S., Wilkinson, C. J., Mayor, T., Mortensen, P., Nigg, E. A., and Mann, M. (2003) Proteomic characterization of the human centrosome by protein correlation profiling. Nature 426, 570-4. (55) Raynaud-Messina, B., and Merdes, A. (2007) Gamma-tubulin complexes and microtubule organization. Curr Opin Cell Biol 19, 24-30. (56) Badano, J. L., Teslovich, T. M., and Katsanis, N. (2005) The centrosome in human genetic disease. Nat Rev Genet 6, 194-205. (57) Burkhard, P., Strelkov, S. V., and Stetefeld, J. (2001) Coiled coils: a highly versatile protein folding motif. Trends Cell Biol. 11, 82-88. (58) Lupas, A. N., and Gruber, M. (2005) The structure of alpha-helical coiled coils. Adv Protein Chem 70, 37-78. (59) Mason, J. M., and Arndt, K. M. (2004) Coiled coil domains: stability, specificity, and biological implications. Chembiochem 5, 170-6. (60) Parry, D. A., Fraser, R. D., and Squire, J. M. (2008) Fifty years of coiled-coils and alpha- helical bundles: a close relationship between sequence and structure. J Struct Biol 163, 258-69. (61) Berger, B., Wilson, D. B., Wolf, E., Tonchev, T., Milla, M., and Kim, P. S. (1995) Predicting coiled coils by use of pairwise residue correlations. Proc. Natl. Acad. Sci. USA 92, 8259-8263. (62) Delorenzi, M., and Speed, T. (2002) An HMM model for coiled-coil domains and a comparison with PSSM-based predictions. Bioinformatics 18, 617-25. (63) Lupas, A., vanDyke, M., and Stock, J. (1991) Predicting coiled coils from protein sequences. Science 252, 1162-1164. (64) McDonnell, A. V., Jiang, T., Keating, A. E., and Berger, B. (2006) Paircoil2: improved prediction of coiled coils from sequence. Bioinformatics 22, 356-8. (65) Liu, J., and Rost, B. (2001) Comparing function and structure between entire proteomes. Protein Sci 10, 1970-9. (66) Rose, A., and Meier, I. (2004) Scaffolds, levers, rods and springs: diverse cellular functions of long coiled-coil proteins. Cell Mol Life Sci 61, 1996-2009. (67) Grigoryan, G., and Keating, A. E. (2008) Structural specificity in coiled-coil interactions. Curr Opin Struct Biol 18, 477-83. (68) Woolfson, D. N. (2005) The design of coiled-coil structures and assemblies. Adv Protein Chem 70, 79-112. (69) Grigoryan, G., Reinke, A. W., and Keating, A. E. (2009) Design of protein-interaction specificity gives selective bZIP-binding peptides. Nature 458, 859-64. (70) Kilmartin, J. V., Dyos, S. L., Kershaw, D., and Finch, J. T. (1993) A spacer protein in the Saccharomyces cerevisiae spindle poly body whose transcript is cell cycle-regulated. J Cell Biol 123, 1175-84. (71) Schaerer, F., Morgan, G., Winey, M., and Philippsen, P. (2001) Cnm67p is a spacer protein of the Saccharomyces cerevisiae spindle pole body outer plaque. Mol Biol Cell 12, 2519-33. (72) Lumb, K. J., Carr, C. M., and Kim, P. S. (1994) Subdomain folding of the coiled coil leucine zipper from the bZIP transcriptional activator GCN4. Biochemistry 33, 7361-7. (73) Oakley, M. G., and Kim, P. S. (1997) Protein dissection of the antiparallel coiled coil from Escherichia coli seryl tRNA synthetase. Biochemistry 36, 2544-9.

(74) Newman, J. R., Wolf, E., and Kim, P. S. (2000) A computationally directed screen identifying interacting coiled coils from Saccharomyces cerevisiae. Proc Natl Acad Sci U S A 97, 13203-8. (75) Zhang, H., Chen, J., Wang, Y., Peng, L., Dong, X., Lu, Y., Keating, A. E., and Jiang, T. (2009) A computationally guided protein-interaction screen uncovers coiled-coil interactions involved in vesicular trafficking. J Mol Biol 392, 228-41. (76) Ishida, H., Takahashi, K., Nakashima, K., Kumaki, Y., Nakata, M., Hikichi, K., and Yazawa, M. (2000) Solution structures of the N-terminal domain of yeast calmodulin: Ca2+-dependent conformational change and its functional implication. Biochemistry 39, 13660-8. (77) Aldaz, H., Rice, L. M., Stearns, T., and Agard, D. A. (2005) Insights into microtubule nucleation from the crystal structure of human gamma-tubulin. Nature 435, 523-7. (78) Hu, H., and Chazin, W. J. (2003) Unique features in the C-terminal domain provide caltractin with target specificity. J Mol Biol 330, 473-84. (79) Wolf, E., Kim, P. S., and Berger, B. (1997) MultiCoil: a program for predicting two- and three-stranded coiled coils. Protein Sci 6, 1179-89. (80) McGuffin, L. J., and Jones, D. T. (2003) Improvement of the GenTHREADER method for genomic fold recognition. Bioinformatics 19, 874-81. (81) Cheng, J. (2007) DOMAC: an accurate, hybrid protein domain prediction server. Nucleic Acids Res 35, W354-6. (82) Zdobnov, E. M., and Apweiler, R. (2001) InterProScan--an integration platform for the signature-recognition methods in InterPro. Bioinformatics 17, 847-8. (83) Hoover, D. M., and Lubkowski, J. (2002) DNAWorks: an automated method for designing oligonucleotides for PCR-based gene synthesis. Nucleic Acids Res 30, e43. (84) Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff, J. W. (1990) Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol 185, 60-89. (85) Edelhoch, H. (1967) Spectroscopic determination of tryptophan and tyrosine in proteins. Biochemistry 6, 1948-1954. (86) Doering, D. S. (1992) Ph.D. thesis in Biology, Massachusetts Institute of Technology, Cambridge, MA. (87) Newman, J. R. S. (2001) Ph.D. thesis in Biology, Massachusetts Institute of Technology, Cambridge, MA. (88) Gordon, G. W., Berry, G., Liang, X. H., Levine, B., and Herman, B. (1998) Quantitative fluorescence resonance energy transfer measurements using fluorescence microscopy. Biophys J 74, 2702-13. (89) Demeler, B. (2005) UltraScan A Comprehensive Data Analysis Software Package for Analytical Ultracentrifugation Experiments, in Modern Analytical Ultracentrifugation: Techniques and Methods (D. J. Scott, S. E. H. a. A. J. R., Ed.) pp 210-229, Royal Society of Chemistry (U.K.). (90) Cole, J. L. a. L., J.W. (2006) Heteroanalysis, Analytical Ultracentrifugation Facility, University of Connecticut, Storrs, CT. (91) Stafford, W. F., and Sherwood, P. J. (2004) Analysis of heterologous interacting systems by sedimentation velocity: curve fitting algorithms for estimation of sedimentation coefficients, equilibrium and kinetic constants. Biophys Chem 108, 231-43. (92) Philo, J. (2006) Sedenterp, Alliance Protein Laboratories, Thousand Oaks, CA.

(93) Otwinowski, Z. a. M., W. (1997) Processing of X-ray Diffraction Data Collected in Oscillation Mode. Methods Enzymol 276, 307-326. (94) (1994) The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol Crystallogr 50, 760-3. (95) O'Shea, E. K., Klemm, J. D., Kim, P. S., and Alber, T. (1991) X-ray structure of the GCN4 leucine zipper, a two-stranded, parallel coiled coil. Science 254, 539-44. (96) Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr 54, 905-21. (97) Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60, 2126-32. (98) Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr A 47 ( Pt 2), 110-9. (99) Strong, M., Sawaya, M. R., Wang, S., Phillips, M., Cascio, D., and Eisenberg, D. (2006) Toward the structural genomics of complexes: crystal structure of a PE/PPE protein complex from Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 103, 8060-5. (100) Winn, M. D., Isupov, M. N., and Murshudov, G. N. (2001) Use of TLS parameters to model anisotropic displacements in macromolecular refinement. Acta Crystallogr D Biol Crystallogr 57, 122-33. (101) ZANUDA (http://www.ysbl.york.ac.uk/YSBLPrograms/) (102) Padilla, J. E., and Yeates, T. O. (2003) A statistic for local intensity differences: robustness to anisotropy and pseudo-centering and utility for detecting twinning. Acta Crystallogr D Biol Crystallogr 59, 1124-30. (103) DeLano, W. L. The Pymol Molecular Graphics System (http://pymol.sourceforge.net) (104) Morris, A. L., MacArthur, M. W., Hutchinson, E. G., and Thornton, J. M. (1992) Stereochemical quality of protein structure coordinates. Proteins 12, 345-64. (105) Kobe, B., and Kajava, A. V. (2001) The leucine-rich repeat as a protein recognition motif. Curr Opin Struct Biol 11, 725-32. (106) Chen, Y. H., Yang, J. T., and Chau, K. H. (1974) Determination of the helix and beta form of proteins in aqueous solution by circular dichroism. Biochemistry 13, 3350-9. (107) Blacklow, S. C., Lu, M., and Kim, P. S. (1995) A trimeric subdomain of the simian immunodeficiency virus envelope glycoprotein. Biochemistry 34, 14955-62. (108) Schnarr, N. A., and Kennan, A. J. (2003) Specific control of peptide assembly with combined hydrophilic and hydrophobic interfaces. J Am Chem Soc 125, 667-71. (109) Crick, F. H. C. (1953) The packing of a-helices: simple coiled coils. Acta Cryst. 6, 689- 697. (110) Harbury, P. B., Tidor, B., and Kim, P. S. (1995) Repacking protein cores with backbone freedom: structure prediction for coiled coils. Proc. Natl. Acad. Sci. U S A 92, 8408- 8412. (111) Walshaw, J., and Woolfson, D. N. (2001) Socket: a program for identifying and analysing coiled-coil motifs within protein structures. J Mol Biol 307, 1427-50.

(112) Vinh, D. B., Kern, J. W., Hancock, W. O., Howard, J., and Davis, T. N. (2002) Reconstitution and characterization of budding yeast gamma-tubulin complex. Mol Biol Cell 13, 1144-57. (113) Chan, D. C., Fass, D., Berger, J. M., and Kim, P. S. (1997) Core structure of gp41 from the HIV envelope glycoprotein. Cell 89, 263-73. (114) Chandramouli, P., Topf, M., Menetret, J. F., Eswar, N., Cannone, J. J., Gutell, R. R., Sali, A., and Akey, C. W. (2008) Structure of the mammalian 80S ribosome at 8.7 A resolution. Structure 16, 535-48. (115) Chial, H. J., Rout, M. P., Giddings, T. H., and Winey, M. (1998) Saccharomyces cerevisiae Ndc1p is a shared component of nuclear pore complexes and spindle pole bodies. J Cell Biol 143, 1789-800. (116) Munoz-Centeno, M. C., McBratney, S., Monterrosa, A., Byers, B., Mann, C., and Winey, M. (1999) Saccharomyces cerevisiae MPS2 encodes a membrane protein localized at the spindle pole body and the nuclear envelope. Mol Biol Cell 10, 2393-406. (117) Schramm, C., Elliott, S., Shevchenko, A., and Schiebel, E. (2000) The Bbp1p-Mps2p complex connects the SPB to the nuclear envelope and is essential for SPB duplication. Embo J 19, 421-33. (118) Winey, M., Goetsch, L., Baum, P., and Byers, B. (1991) MPS1 and MPS2: novel yeast genes defining distinct steps of spindle pole body duplication. J Cell Biol 114, 745-54. (119) Winey, M., Hoyt, M. A., Chan, C., Goetsch, L., Botstein, D., and Byers, B. (1993) NDC1: a nuclear periphery component required for yeast spindle pole body duplication. J Cell Biol 122, 743-51. (120) Biggins, S., and Rose, M. D. (1994) Direct interaction between yeast spindle pole body components: Kar1p is required for Cdc31p localization to the spindle pole body. J Cell Biol 125, 843-52. (121) Jaspersen, S. L., Giddings, T. H., Jr., and Winey, M. (2002) Mps3p is a novel component of the yeast spindle pole body that interacts with the yeast centrin homologue Cdc31p. J Cell Biol 159, 945-56. (122) Kilmartin, J. V. (2003) Sfi1p has conserved centrin-binding sites and an essential function in budding yeast spindle pole body duplication. J Cell Biol 162, 1211-21. (123) Spang, A., Courtney, I., Fackler, U., Matzner, M., and Schiebel, E. (1993) The calcium- binding protein cell division cycle 31 of Saccharomyces cerevisiae is a component of the half bridge of the spindle pole body. J Cell Biol 123, 405-16. (124) Spang, A., Courtney, I., Grein, K., Matzner, M., and Schiebel, E. (1995) The Cdc31p- binding protein Kar1p is a component of the half bridge of the yeast spindle pole body. J Cell Biol 128, 863-77. (125) Vallen, E. A., Ho, W., Winey, M., and Rose, M. D. (1994) Genetic interactions between CDC31 and KAR1, two genes required for duplication of the microtubule organizing center in Saccharomyces cerevisiae. Genetics 137, 407-22. (126) Rout, M. P., and Kilmartin, J. V. (1991) Yeast spindle pole body components. Cold Spring Harb Symp Quant Biol 56, 687-92. (127) Elliott, S., Knop, M., Schlenstedt, G., and Schiebel, E. (1999) Spc29p is a component of the Spc110p subcomplex and is essential for spindle pole body duplication. Proc Natl Acad Sci U S A 96, 6205-10.

(128) Knop, M., and Schiebel, E. (1997) Spc98p and Spc97p of the yeast gamma-tubulin complex mediate binding to the spindle pole body via their interaction with Spc110p. Embo J 16, 6985-95. (129) Zizlsperger, N., Malashkevich, V. N., Pillay, S., and Keating, A. E. (2008) Analysis of coiled-coil interactions between core proteins of the spindle pole body. Biochemistry 47, 11858-68. (130) Soding, J., Biegert, A., and Lupas, A. N. (2005) The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res 33, W244-8. (131) Newman, J. R., and Keating, A. E. (2003) Comprehensive identification of human bZIP interactions with coiled-coil arrays. Science 300, 2097-101. (132) Kippert, F., and Gerloff, D. L. (2009) Highly sensitive detection of individual HEAT and ARM repeats with HHpred and COACH. PLoS One 4, e7148. (133) Jaspersen, S. L., Martin, A. E., Glazko, G., Giddings, T. H., Jr., Morgan, G., Mushegian, A., and Winey, M. (2006) The Sad1-UNC-84 homology domain in Mps3 interacts with Mps2 to connect the spindle pole body with the nuclear envelope. J Cell Biol 174, 665- 75. (134) Donaldson, A. D., and Kilmartin, J. V. (1996) Spc42p: a phosphorylated component of the S. cerevisiae spindle pole body (SPD) with an essential function during SPB duplication. J Cell Biol 132, 887-901. (135) Nguyen, T., Vinh, D. B., Crawford, D. K., and Davis, T. N. (1998) A genetic analysis of interactions with Spc110p reveals distinct functions of Spc97p and Spc98p, components of the yeast gamma-tubulin complex. Mol Biol Cell 9, 2201-16. (136) Pereira, G., Knop, M., and Schiebel, E. (1998) Spc98p directs the yeast gamma-tubulin complex into the nucleus and is subject to cell cycle-dependent phosphorylation on the nuclear side of the spindle pole body. Mol Biol Cell 9, 775-93. (137) Spang, A., Geissler, S., Grein, K., and Schiebel, E. (1996) gamma-Tubulin-like Tub4p of Saccharomyces cerevisiae is associated with the spindle pole body substructures that organize microtubules and is required for mitotic spindle formation. J Cell Biol 134, 429- 41. (138) Moritz, M., Braunfeld, M. B., Sedat, J. W., Alberts, B., and Agard, D. A. (1995) Microtubule nucleation by gamma-tubulin-containing rings in the centrosome. Nature 378, 638-40. (139) Zheng, Y., Wong, M. L., Alberts, B., and Mitchison, T. (1995) Nucleation of microtubule assembly by a gamma-tubulin-containing ring complex. Nature 378, 578-83. (140) Jaspersen, S. L., and Stearns, T. (2008) Exploring the pole: an EMBO conference on centrosomes and spindle pole bodies. Nat Cell Biol 10, 1375-8. (141) Geiser, J. R., Sundberg, H. A., Chang, B. H., Muller, E. G., and Davis, T. N. (1993) The essential mitotic target of calmodulin is the 110-kilodalton component of the spindle pole body in Saccharomyces cerevisiae. Mol Cell Biol 13, 7913-24. (142) Sundberg, H. A., and Davis, T. N. (1997) A mutational analysis identifies three functional regions of the spindle pole component Spc110p in Saccharomyces cerevisiae. Mol Biol Cell 8, 2575-90. (143) Brown, J. H., Cohen, C., and Parry, D. A. (1996) Heptad breaks in alpha-helical coiled coils: stutters and stammers. Proteins 26, 134-45. (144) Schwaiger, I., Sattler, C., Hostetter, D. R., and Rief, M. (2002) The myosin coiled-coil is a truly elastic protein structure. Nat Mater 1, 232-5.

100

(145) Stirling, D. A., Welch, K. A., and Stark, M. J. (1994) Interaction with calmodulin is required for the function of Spc110p, an essential component of the yeast spindle pole body. Embo J 13, 4329-42. (146) Yamniuk, A. P., and Vogel, H. J. (2004) Calmodulin's flexibility allows for promiscuity in its interactions with target proteins and peptides. Mol Biotechnol 27, 33-57. (147) Gruneberg, U., Campbell, K., Simpson, C., Grindlay, J., and Schiebel, E. (2000) Nud1p links astral microtubule organization and the control of exit from mitosis. Embo J 19, 6475-88. (148) Usui, T., Maekawa, H., Pereira, G., and Schiebel, E. (2003) The XMAP215 homologue Stu2 at yeast spindle pole bodies regulates microtubule dynamics and anchorage. Embo J 22, 4779-93. (149) Le Masson, I., Saveanu, C., Chevalier, A., Namane, A., Gobin, R., Fromont-Racine, M., Jacquier, A., and Mann, C. (2002) Spc24 interacts with Mps2 and is required for chromosome segregation, but is not implicated in spindle pole body duplication. Mol Microbiol 43, 1431-43. (150) Geier, B. M., Wiech, H., and Schiebel, E. (1996) Binding of centrins and yeast calmodulin to synthetic peptides corresponding to binding sites in the spindle pole body components Kar1p and Spc110p. J Biol Chem 271, 28366-74. (151) Keskin, O., Tuncbag, N., and Gursoy, A. (2008) Characterization and prediction of protein interfaces to infer protein-protein interaction networks. Curr Pharm Biotechnol 9, 67-76. (152) Zhou, H. X., and Qin, S. (2007) Interaction-site prediction for protein complexes: a critical assessment. Bioinformatics 23, 2203-9. (153) Grigoryan, G., and Keating, A. E. (2006) Structure-based prediction of bZIP partnering specificity. J Mol Biol 355, 1125-42. (154) Greenfield, N. J., Montelione, G. T., Farid, R. S., and Hitchcock-DeGregori, S. E. (1998) The structure of the N-terminus of striated muscle alpha-tropomyosin in a chimeric peptide: nuclear magnetic resonance structure and circular dichroism studies. Biochemistry 37, 7834-43. (155) Weissenhorn, W., Dessen, A., Harrison, S. C., Skehel, J. J., and Wiley, D. C. (1997) Atomic structure of the ectodomain from HIV-1 gp41. Nature 387, 426-30. (156) Castillo, A. R., Meehl, J. B., Morgan, G., Schutz-Geschwender, A., and Winey, M. (2002) The yeast protein kinase Mps1p is required for assembly of the integral spindle pole body component Spc42p. J Cell Biol 156, 453-65. (157) Friedman, D. B., Kern, J. W., Huneycutt, B. J., Vinh, D. B., Crawford, D. K., Steiner, E., Scheiltz, D., Yates, J., 3rd, Resing, K. A., Ahn, N. G., Winey, M., and Davis, T. N. (2001) Yeast Mps1p phosphorylates the spindle pole component Spc110p in the N- terminal domain. J Biol Chem 276, 17958-67. (158) Holinger, E. P., Old, W. M., Giddings, T. H., Jr., Wong, C., Yates, J. R., 3rd, and Winey, M. (2009) Budding yeast centrosome duplication requires stabilization of Spc29 via Mps1-mediated phosphorylation. J Biol Chem 284, 12949-55. (159) Jaspersen, S. L., Huneycutt, B. J., Giddings, T. H., Jr., Resing, K. A., Ahn, N. G., and Winey, M. (2004) Cdc28/Cdk1 regulates spindle pole body duplication through phosphorylation of Spc42 and Mps1. Dev Cell 7, 263-74.

101

(160) Singh, A., and Hitchcock-DeGregori, S. E. (2003) Local destabilization of the tropomyosin coiled coil gives the molecular flexibility required for actin binding. Biochemistry 42, 14114-21.

102

APPENDIX

Predicted coiled-coil regions of SPB protein – DNA and protein sequences. Residue numbers are given in Table 1. Peptides were expressed with the following extra tags and linkers:

N-terminal cysteine: SYKCGGS-coiledcoil-EFGDYKDDDDKGHHHHHH

C-terminal cysteine: SYKGS-coiledcoil-EFGGCGGDYKDDDDKGHHHHHH

coiled-coil systematic Paircoil2 coiled-coil DNA sequence coiled-coil protein sequence name name P score

AAAATCTTCCTGCAGAACAGCCTGTCTAAAGAAGACTTCCGTATGCTCGAG KIFLQNSLSKEDFRMLENVI AATGTAATCCTGGGTTACCAGAAAAAAGTTATCGAACTGGGTCGTGACAAC LGYQKKVIELGRDNLRQEE CTGCGTCAGGAGGAACGTGCGAACTCTCTCCAGAAGGAACTGGAAGCGGCG RANSLQKELEAATKSNDKT Cnm67_1 YNL225c 1.00 ACCAAATCTAACGACAAAACCCTGGACAACAAAAAGAAAATCGAAGAACA LDNKKKIEEQTVLIENLTKD GACCGTTCTGATCGAGAACCTGACCAAAGACCTGTCCCTGAACAAAGAGAT LSLNKEMLEKANDTIQTKH GCTGGAAAAAGCGAACGACACCATCCAGACCAAACACACCGCGCTGCTGTC TALLSLTDSLRKAELFEI TCTGACCGATTCTCTGCGTAAAGCAGAACTGTTCGAAATT

ATTAAAGGTTTCCTGGCAGCGTCTCAGCAGGAAGAACTGAGCCGTATCTCT IKGFLAASQQEELSRISQRF CAGCGTTTCAAAAACGCGAAAGCGGAAGCAGAAGACCTGCGCAACGAACT KNAKAEAEDLRNELENKKI GGAAAACAAGAAAATCGAAATCCAGACCATGCGTGAAAAGAACAACACCC EIQTMREKNNTLIGTNKTLS TGATCGGTACTAACAAAACCCTGAGCAAACAGAACAAAATCCTCGCTGACA KQNKILADKFDKLTIDEKEI AATTCGACAAACTCACTATCGACGAAAAAGAAATCCTGAAAGGTGCGAACG Cnm67_2 YNL225c 1.00 LKGANEEIKIKLERLNERLG AAGAGATCAAGATTAAACTCGAACGCCTGAATGAGCGCCTGGGTTCTTGGG SWEKSKEKYETSLKDKEK AGAAGAGCAAAGAAAAGTACGAAACCAGCCTCAAAGATAAAGAGAAAATG MLADAEKKTNTLSKELDN CTGGCTGACGCGGAGAAAAAAACTAACACTCTGTCTAAGGAGCTGGATAAC LRSRFGNLEGNTSERITIKNI CTGCGTTCTCGTTTCGGTAACCTGGAAGGTAACACCTCTGAACGTATCACCA LQS TTAAGAACATTCTGCAGTCT ATCAAAGGTTTCCTGGCGGCGTCTCAGCAGGAAGAACTGTCTCGTATCTCTC IKGFLAASQQEELSRISQRF AGCGTTTCAAAAACGCGAAAGCTGAAGCGGAAGACCTGCGTAACGAACTG KNAKAEAEDLRNELENKKI Cnm67_3 YNL225c 1.00 GAAAACAAAAAGATCGAAATCCAGACCATGCGTGAAAAAAACAACACCCT EIQTMREKNNTLIGTNKTLS GATCGGTACCAACAAAACCCTGTCTAAACAGAACAAAATTCTGGCGGACAA KQNKILADKFDKLTIDEKEI ATTCGACAAACTGACCATCGACGAAAAAGAAATCCTGAAAGGT LKG

103

GATAAATTCGACAAACTGACTATCGACGAAAAAGAGATCCTGAAGGGTGCG DKFDKLTIDEKEILKGANEE AACGAAGAAATCAAGATCAAACTGGAACGTCTGAACGAACGCCTGGGTTCT IKIKLERLNERLGSWEKSKE TGGGAAAAATCTAAGGAGAAATACGAAACCAGCCTGAAAGACAAAGAGAA Cnm67_4 YNL225c 1.00 KYETSLKDKEKMLADAEK AATGCTGGCTGACGCGGAAAAGAAAACTAACACTCTGTCCAAAGAACTGGA KTNTLSKELDNLRSRFGNL CAACCTGCGTTCTCGTTTCGGTAACCTGGAAGGCAACACTTCCGAACGCATC EGNTSERITIKNILQS ACCATCAAGAACATCCTGCAGTCT GGTAACTCTGCGTCTAAAAAGTTCCAGGACGACACCCTGAACCGTGTTCGT GNSASKKFQDDTLNRVRKE Spc29_1 YPL124w 0.99 AAAGAACACGAAGAAGCGCTGAAAAAACTGCGTGAAGAAAACTTCTCTTCT HEEALKKLREENFSSNTSEL AACACCTCTGAACTC GCGTCTCAGAACGTTATCGACGATCAGCGCCTGGAAATCAAATACCTGGAA ASQNVIDDQRLEIKYLERIV Spc29_2 YPL124w 1.00 CGTATCGTTTACGACCAGGGTACCGTTATTGACAACCTGACCTCTCGTATCA YDQGTVIDNLTSRITRLESFI CCCGTCTGGAATCTTTCATCCTGAACTCTATCTCT LNSIS GAAGAATACAAACGTAACACCGAATTTATCAACAAAGCGGTGCAGCAGAA EEYKRNTEFINKAVQQNKE CAAAGAGCTGAACTTCAAACTGCGTGAAAAACAGAACGAAATCTTCGAACT LNFKLREKQNEIFELKKIAE GAAAAAGATCGCGGAAACCCTGCGCTCCAAACTGGAAAAATACGTTGACAT Spc42_1 YKL042w 1.00 TLRSKLEKYVDITKKLEDQ CACCAAAAAGCTGGAAGACCAGAACCTGAACCTGCAGATCAAAATCTCTGA NLNLQIKISDLEKKLSDANS CCTGGAGAAAAAACTGAGCGACGCGAACTCTACCTTCAAAGAAATGCGTTT TFKEMRF C TCGTCTGAAAGCGATCGAACGTACCCTGTCCGTGCTGACCAACTACGTGAT SNTSDQDSRLKAIERTLSVL Spc42_2 YKL042w 0.99 GCGTTCTGAAGACGGTAACAACGACCGCATGTCT TNYVMRSEDGNNDRMS TCTGACGACGATATTATGATGTACGAATCTGCGGAACTGAAACGCGTGGAA SDDDIMMYESAELKRVEEE GAAGAAATCGAAGAACTCAAACGCAAAATTCTGGTTCGTAAAAAACACGA IEELKRKILVRKKHDLRKLS Spc42_3 YKL042w 0.99 CCTGCGTAAGCTGTCTCTGAACAACCAGCTGCAGGAACTCCAGTCTATGAT LNNQLQELQSMMDGDDNI GGACGGTGACGACAACATCAAACTGGACAACGTTAGCAAACACAACCACG KLDNVSKHNHATH CGACCCAC TCTCTGGGTAACGATACCGATTTCCGTAACTCTATCATCGAAGGTCTGAACC SLGNDTDFRNSIIEGLNLEIN TGGAAATCAACAAACTGAAACAGGATCTGAAAGCGAAAGAAGTTGAATAC KLKQDLKAKEVEYQDTLQ Spc72_1 YAL047c 1.00 CAGGACACCCTGCAGTTCGTTCAGGAAAACCTCGAAAACTCTGAATCTATC FVQENLENSESIVNTINHLL GTGAACACCATTAACCACCTGCTGTCCTTCATCCTGACT SFILT

104

GAATACGACCAGTTCATTAACTCTATCCGTCTGAAGTTCGAAAAAAGCCAG AAACTCGAGAAAATCATCGCGTCTAAACTGAACGAACAGTCTCACCTGCTG EYDQFINSIRLKFEKSQKLE GACTCTCTGGAGCTGGAAGAAAACTCCTCTAGCGTTATTGAAAAACAGGAC KIIASKLNEQSHLLDSLELE CACCTGATTTCCCAGCTCAAAGAAAAGATCGAATCCCAGTCTGTACTGATTA ENSSSVIEKQDHLISQLKEKI ACAACCTGGAAAAGCTGAAAGAGGACATCATCAAGATGAAACAGAACGAA ESQSVLINNLEKLKEDIIKM Spc72_2 YAL047c 1.00 AAGGTTCTCACCAAAGAACTGGAAACCCAGACCAAAATCAACAAACTCAA KQNEKVLTKELETQTKINK GGAGAACAACTGGGACTCTTACATTAACGACCTGGAGAAGCAGATCAATGA LKENNWDSYINDLEKQIND CCTCCAGATCGACAAATCTGAAGAGTTCCACGTTATTCAGAATCAACTCGA LQIDKSEEFHVIQNQLDKLD CAAGCTCGACCTCGAAAACTATCAACTGAAGAACCAACTCAACACCCTGGA LENYQLKNQLNTLDNQKLI TAACCAGAAGCTGATCCTCAGCCAATACGAATCCAACTTCATCAAATTCAA LSQYESNFIKFNQNLL CCAAAACCTGCTC AACAAAGAGCTGACCCTGCGTATCGAAGAACTGCAGCGTCGTTGGATTTCT NKELTLRIEELQRRWISERE GAACGTGAACGTCGTAAACTGGACGCGAACGCGTCTGAAGCGCGTATCAAA Spc72_3 YAL047c 1.00 RRKLDANASEARIKALEQE GCGCTGGAACAGGAAAACGAATCTCTGCGTTCTAAACTGTTCAACCTGTCT NESLRSKLFNLSINNP ATCAACAACCCG CGTGAACTGGAACAGATCATCAACGAAACCGAAGTTAACAAACAGATGGA RELEQIINETEVNKQMELLY ACTGCTGTACAACATCTACGAAGAAATCTTCCGTGAAATCGAAGAACGTCG NIYEEIFREIEERRTNQSSQE Spc97 YHR172w 0.98 CACCAACCAGTCTAGCCAGGAAGACTTCAACAACTTCATGGACAGCATGAA DFNNFMDSMKNESSLHRL AAACGAGTCTTCTCTGCACCGTCTGATGGTTGCGTTCGAT MVAFD CCGCTGATCCGTGACATCATCAACAAACTGTCTCGTATCTCTATCCTGCGTA PLIRDIINKLSRISILRTQFQQ CCCAGTTCCAGCAGTTCAACTCTAAGATGGAAAGCTACTACCTGAATaGCAT FNSKMESYYLNCIIEENFKE Spc98_1 YNL126w 0.97 CATCGAGGAGAACTTCAAAGAAATGACCCGTAAACTGCAGCGTACCGAGA MTRKLQRTENKSQNQFDLI ACAAATCTCAGAACCAGTTCGACCTCATCCGTCTGAACAACGGTACCATCG RLNNGTIE AA AAGATGAATCTGAACGACCACGAAGCGTCTAACGGTCTGCTGGGTAAATTC KMNLNDHEASNGLLGKFN Spc98_2 YNL126w 0.97 AACACCAACCTGAAGGAGATCGTTAGCCAATACAAAAACTTCAAGGATCGT TNLKEIVSQYKNFKDRLYIF CTGTACATCTTCCGTGCGGACCTGAAAAAT RADLKN CACAAAAACCGTGAATACAAAAAAGCGTACTTCGATCTGTTCGCGCAGATG HKNREYKKAYFDLFAQMD GATCTGAACTCCCGTGACCTGGAAGATCTGGCGGAAGATGTTCGTGAGCAA LNSRDLEDLCEDVREQREQ CGCGAGCAGTTCCACCGTAATGAGCAGACTTACAAACAGGCGTACGAAGAA FHRNEQTYKQAYEEMRAE ATGCGTGCGGAACTGGTTAACGAGCTGAAAAAATCCAAAACCCTGTTCGAA LVNELKKSKTLFENYYSLG AACTACTACAGCCTCGGTCAGAAATACAAGTCTCTGAAGAAAGTACTGGAC Bbp1_1 YPL255W 1.00 QKYKSLKKVLDQTISHEAE CAGACCATCTCCCACGAAGCAGAACTGGCGACCTCTCGTGAACGCCTCTAT LATSRERLYQEEDLKNFEIQ CAAGAGGAAGACCTCAAAAACTTCGAAATCCAGACCCTGAAACAGCGTCTG TLKQRLSDLELKYTNLQIE TCTGACCTCGAACTCAAGTACACCAACCTGCAGATTGAAAAAGACATGCAG KDMQRDNYESEIHDLLLQL CGTGACAACTACGAAAGCGAAATTCACGACCTGCTGCTCCAGCTGTCCCTG SLRNNERKDTSAGSN CGTAACAACGAACGTAAAGACACCTCTGCTGGTTCTAAC

105

CACAAAAACCGTGAATACAAGAAAGCGTACTTCGATCTGTTTGCGCAGATG HKNREYKKAYFDLFAQMD GACCTGAACTCTCGTGACCTGGAAGATCTGGCGGAAGATGTACGTGAGCAG LNSRDLEDLCEDVREQREQ Bbp1_2 YPL255W 1.00 CGTGAACAGTTCCACCGTAACGAACAGACCTACAAACAGGCGTACGAAGA FHRNEQTYKQAYEEMRAE AATGCGTGCGGAACTGGTGAACGAACTGAAGAAGTCTAAAACCCTGTTCGA LVNELKKSKTLFENYYSLG AAACTACTACTCTCTGGGTCAGAAATACAAA QKYK TCCCTGGGTCAGAAATACAAATCTCTGAAGAAGGTTCTGGACCAGACCATC SLGQKYKSLKKVLDQTISH TCTCACGAAGCGGAACTGGCGACCTCTCGTGAACGTCTGTACCAGGAAGAA EAELATSRERLYQEEDLKN GACCTGAAAAACTTCGAAATCCAGACCCTGAAACAGCGTCTGTCTGACCTG Bbp1_3 YPL255W 1.00 FEIQTLKQRLSDLELKYTNL GAACTGAAGTACACCAACCTGCAGATCGAAAAAGACATGCAGCGTGACAA QIEKDMQRDNYESEIHDLL CTACGAATCCGAGATCCACGACCTGCTCCTGCAGCTGTCTCTGCGTAACAAC LQLSLRNNERKDTSAGSN GAACGTAAAGACACCTCTGCGGGTTCTAAC ATCAAACTGCTGTCTCGCAACAACATCGGCAAAGCACTGGAGGTGCAGGTG IKLLSRNNIGKALEVQVEEL GAAGAACTGAAACGTGAACTGACCGCGAAACAGTCCCTGCTGCAGGAAAA KRELTAKQSLLQENERQVS Mps2_1 YGL075C 1.00 CGAACGCCAGGTTAGCGAGCTGAAAATCCGTCTGGAAACCTACCAGGAGAA ELKIRLETYQEKYASIQQRF ATACGCTTCTATCCAGCAGCGTTTCTCCGACCTGCAGAAAGCTCGTCAGGTT SDLQKARQVEDNQNSSRTS GAAGACAACCAGAACTCTTCCCGTACCTCC GTGACCGGCATCGATCAGAAAGCGATTCTGGAAGAGTTCCGTCGTCGTCTG VTGIDQKAILEEFRRRLQRQ Mps2_2 YGL075C 1.00 CAGCGTCAGACCGATACCATCTCTTTCCTGAAAGACCAGATCCGTCGCGAA TDTISFLKDQIRRERGLNCS CGTGGTCTGAACTcCTCCAACGACAAA NDK CGTAACCAGTCTCTGTACCTGGATCGTGAAATCCTGCTGCAACGTCAGATCA RNQSLYLDREILLQRQIKKR AAAAACGTGACGAAAAAATCAAAGCGCTGGAATCTAAACTGCAATCTCTGC Nbp1 YLR457C 1.00 DEKIKALESKLQSLQEALN AGGAAGCTCTGAACTACTCTAACGAAAAGTACCGTATCCTGGAAGACCTGC YSNEKYRILEDLLDSSNI TGGACTCTTCTAACATC CTGATTCTGAACGAAGCGCTGAAAACCATCCAGATCAACAACGAAAAAGTT LILNECLKTIQINNEKVVQY Ndc1 YML031W 0.98 GTTCAGTACCTGCGTTCTGTTCAGGATCTGGGTGGTTCTGCGACCGCGCGTC LRSVQDLGGSATARHK ACAAA GACGCGTTCGACTACAACGAAGGTATCGCGTCTCGTACCAAAAACATCAAC DAFDYNEGIASRTKNINSDS Kar1_1 YNL188W 0.97 TCCGACTCTGACCGTTCTAACGACACTATCAAG DRSNDTIK CTGGCGGAAAACAAAGCGGAAGAATACATCTCTGACGAGGACAACGTTAA LAENKAEEYISDEDNVKID Kar1_2 YNL188W 0.98 AATCGACGAAGACAACATCGAAAACGAACTGCAA EDNIENELQ ATCCTGCAGTCTGAAATCGAAATGCACACCAAAAAACTGGACACCATCATC ILQSEIEMHTKKLDTIIELLK Kar1_3 YNL188W 0.98 GAACTGCTGAAAGACGACACCGACTCTAAAGAA DDTDSKE

106

AAATCTTTCTCCAACCTGCAGAAACAGGTTAATCACCTGTACTCTGAACTGT KSFSNLQKQVNHLYSELSK CTAAACGTGACGAAAAACACTCTAGCGAACTGGACAAAACCGTTAAAATCA RDEKHSSELDKTVKIIVSQF TCGTTTCTCAGTTCGAAAAAAACATCAAACGTCTGCTGCCGTCTAACCTGGT Mps3_1 YJL019W 1.00 EKNIKRLLPSNLVNFENDIN TAACTTTGAAAACGACATCAACTCTCTGACCAAACAAGTTGAAACCATCTCT SLTKQVETISTSMSELQRRN ACCTCTATGTCCGAACTCCAGCGTCGTAACCACAAATTCACCGTTGAAAATG HKFTVENVTQWQ TTACCCAGTGGCAA AAATCTTTCTCTAACCTGCAGAAACAAGTTAACCACCTGTACTCTGAACTGT KSFSNLQKQVNHLYSELSK Mps3_2 YJL019W 1.00 CTAAACGTGACGAAAAACACTCTTCTGAGCTGGACAAGACCGTTAAAATCA RDEKHSSELDKTVKIIVSQF TCGTTTCTCAGTTCGAAAAAAAC EKN AAAATCATCGTTTCTCAGTTCGAAAAAAACATCAAACGTCTGCTGCCGTCTA KIIVSQFEKNIKRLLPSNLV ACCTGGTGAACTTCGAGAACGACATCAACTCTCTGACCAAACAGGTTGAAA NFENDINSLTKQVETISTSM Mps3_3 YJL019W 1.00 CCATCTCTACCTCTATGTCTGAACTGCAGCGTCGTAACCACAAATTCACCGT SELQRRNHKFTVENVTQW TGAAAACGTTACTCAGTGGCAA Q AAAGAAATTCTGTCTAACGAACTCCAGTACATCGACAAAGACTACTTCATC KEILSNELQYIDKDYFIQEM Mps3_4 YJL019W 0.99 CAGGAAATGAACCGTCGTCTGCAGTCTAACAAACAGGAGATCTGGGAAGA NRRLQSNKQEIWEEITNRLE AATCACCAACCGTCTGGAAACCCAGCAGCAA TQQQ

107